Upgrade to OpenAPI 3.1.0 with prefixItems and streamlined validation

- Upgraded OpenAPI spec from 3.0.3 to 3.1.0
- Replaced QueueItem oneOf pattern with prefixItems for precise tuple validation
- Simplified GitHub Actions workflow to use only Python tests (removed redundant swagger-editor validation)
- All validation now handled by openapi-spec-validator which supports OpenAPI 3.1
- QueueItem now enforces exact tuple structure: [position, prompt_id, prompt, extra_data, outputs_to_execute]

.github/workflows/openapi-validation.yml

@@ -0,0 +1,33 @@
+name: Validate OpenAPI
+
+on:
+  push:
+    branches: [ master ]
+    paths:
+      - 'openapi.yaml'
+      - 'tests-api/openapi.yaml'
+  pull_request:
+    branches: [ master ]
+    paths:
+      - 'openapi.yaml'
+      - 'tests-api/openapi.yaml'
+
+jobs:
+  validate:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+
+      - name: Set up Python
+        uses: actions/setup-python@v4
+        with:
+          python-version: '3.11'
+
+      - name: Install test dependencies
+        run: |
+          pip install -r tests-api/requirements.txt
+
+      - name: Run OpenAPI spec validation tests
+        run: |
+          pytest tests-api/test_spec_validation.py -v

.gitignore

@@ -21,6 +21,5 @@ venv/
 *.log
 web_custom_versions/
 .DS_Store
-openapi.yaml
 filtered-openapi.yaml
 uv.lock

README.md

@@ -110,7 +110,6 @@ ComfyUI follows a weekly release cycle every Friday, with three interconnected r
 
 2. **[ComfyUI Desktop](https://github.com/Comfy-Org/desktop)**
    - Builds a new release using the latest stable core version
-   - Version numbers match the core release (e.g., Desktop v1.7.0 uses Core v1.7.0)
 
 3. **[ComfyUI Frontend](https://github.com/Comfy-Org/ComfyUI_frontend)**
    - Weekly frontend updates are merged into the core repository
@@ -198,11 +197,11 @@ Put your VAE in: models/vae
 ### AMD GPUs (Linux only)
 AMD users can install rocm and pytorch with pip if you don't have it already installed, this is the command to install the stable version:
 
-```pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/rocm6.2.4```
+```pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/rocm6.3```
 
-This is the command to install the nightly with ROCm 6.3 which might have some performance improvements:
+This is the command to install the nightly with ROCm 6.4 which might have some performance improvements:
 
-```pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/rocm6.3```
+```pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/rocm6.4```
 
 ### Intel GPUs (Windows and Linux)
 
@@ -302,7 +301,7 @@ For AMD 7600 and maybe other RDNA3 cards: ```HSA_OVERRIDE_GFX_VERSION=11.0.0 pyt
 
 ### AMD ROCm Tips
 
-You can enable experimental memory efficient attention on pytorch 2.5 in ComfyUI on RDNA3 and potentially other AMD GPUs using this command:
+You can enable experimental memory efficient attention on recent pytorch in ComfyUI on some AMD GPUs using this command, it should already be enabled by default on RDNA3. If this improves speed for you on latest pytorch on your GPU please report it so that I can enable it by default.
 
 ```TORCH_ROCM_AOTRITON_ENABLE_EXPERIMENTAL=1 python main.py --use-pytorch-cross-attention```
 

comfy/comfy_types/node_typing.py

@@ -235,7 +235,7 @@ class ComfyNodeABC(ABC):
     DEPRECATED: bool
     """Flags a node as deprecated, indicating to users that they should find alternatives to this node."""
     API_NODE: Optional[bool]
-    """Flags a node as an API node."""
+    """Flags a node as an API node. See: https://docs.comfy.org/tutorials/api-nodes/overview."""
 
     @classmethod
     @abstractmethod

comfy/ldm/hunyuan_video/model.py

@@ -228,6 +228,7 @@ class HunyuanVideo(nn.Module):
         y: Tensor,
         guidance: Tensor = None,
         guiding_frame_index=None,
+        ref_latent=None,
         control=None,
         transformer_options={},
     ) -> Tensor:
@@ -238,6 +239,14 @@ class HunyuanVideo(nn.Module):
         img = self.img_in(img)
         vec = self.time_in(timestep_embedding(timesteps, 256, time_factor=1.0).to(img.dtype))
 
+        if ref_latent is not None:
+            ref_latent_ids = self.img_ids(ref_latent)
+            ref_latent = self.img_in(ref_latent)
+            img = torch.cat([ref_latent, img], dim=-2)
+            ref_latent_ids[..., 0] = -1
+            ref_latent_ids[..., 2] += (initial_shape[-1] // self.patch_size[-1])
+            img_ids = torch.cat([ref_latent_ids, img_ids], dim=-2)
+
         if guiding_frame_index is not None:
             token_replace_vec = self.time_in(timestep_embedding(guiding_frame_index, 256, time_factor=1.0))
             vec_ = self.vector_in(y[:, :self.params.vec_in_dim])
@@ -313,6 +322,8 @@ class HunyuanVideo(nn.Module):
                         img[:, : img_len] += add
 
         img = img[:, : img_len]
+        if ref_latent is not None:
+            img = img[:, ref_latent.shape[1]:]
 
         img = self.final_layer(img, vec, modulation_dims=modulation_dims)  # (N, T, patch_size ** 2 * out_channels)
 
@@ -324,7 +335,7 @@ class HunyuanVideo(nn.Module):
         img = img.reshape(initial_shape[0], self.out_channels, initial_shape[2], initial_shape[3], initial_shape[4])
         return img
 
-    def forward(self, x, timestep, context, y, guidance=None, attention_mask=None, guiding_frame_index=None, control=None, transformer_options={}, **kwargs):
+    def img_ids(self, x):
         bs, c, t, h, w = x.shape
         patch_size = self.patch_size
         t_len = ((t + (patch_size[0] // 2)) // patch_size[0])
@@ -334,7 +345,11 @@ class HunyuanVideo(nn.Module):
         img_ids[:, :, :, 0] = img_ids[:, :, :, 0] + torch.linspace(0, t_len - 1, steps=t_len, device=x.device, dtype=x.dtype).reshape(-1, 1, 1)
         img_ids[:, :, :, 1] = img_ids[:, :, :, 1] + torch.linspace(0, h_len - 1, steps=h_len, device=x.device, dtype=x.dtype).reshape(1, -1, 1)
         img_ids[:, :, :, 2] = img_ids[:, :, :, 2] + torch.linspace(0, w_len - 1, steps=w_len, device=x.device, dtype=x.dtype).reshape(1, 1, -1)
-        img_ids = repeat(img_ids, "t h w c -> b (t h w) c", b=bs)
+        return repeat(img_ids, "t h w c -> b (t h w) c", b=bs)
+
+    def forward(self, x, timestep, context, y, guidance=None, attention_mask=None, guiding_frame_index=None, ref_latent=None, control=None, transformer_options={}, **kwargs):
+        bs, c, t, h, w = x.shape
+        img_ids = self.img_ids(x)
         txt_ids = torch.zeros((bs, context.shape[1], 3), device=x.device, dtype=x.dtype)
-        out = self.forward_orig(x, img_ids, context, txt_ids, attention_mask, timestep, y, guidance, guiding_frame_index, control, transformer_options)
+        out = self.forward_orig(x, img_ids, context, txt_ids, attention_mask, timestep, y, guidance, guiding_frame_index, ref_latent, control=control, transformer_options=transformer_options)
         return out

comfy/ldm/wan/model.py

@@ -247,6 +247,60 @@ class VaceWanAttentionBlock(WanAttentionBlock):
         return c_skip, c
 
 
+class WanCamAdapter(nn.Module):
+    def __init__(self, in_dim, out_dim, kernel_size, stride, num_residual_blocks=1, operation_settings={}):
+        super(WanCamAdapter, self).__init__()
+
+        # Pixel Unshuffle: reduce spatial dimensions by a factor of 8
+        self.pixel_unshuffle = nn.PixelUnshuffle(downscale_factor=8)
+
+        # Convolution: reduce spatial dimensions by a factor
+        #  of 2 (without overlap)
+        self.conv = operation_settings.get("operations").Conv2d(in_dim * 64, out_dim, kernel_size=kernel_size, stride=stride, padding=0, device=operation_settings.get("device"), dtype=operation_settings.get("dtype"))
+
+        # Residual blocks for feature extraction
+        self.residual_blocks = nn.Sequential(
+            *[WanCamResidualBlock(out_dim, operation_settings = operation_settings) for _ in range(num_residual_blocks)]
+        )
+
+    def forward(self, x):
+        # Reshape to merge the frame dimension into batch
+        bs, c, f, h, w = x.size()
+        x = x.permute(0, 2, 1, 3, 4).contiguous().view(bs * f, c, h, w)
+
+        # Pixel Unshuffle operation
+        x_unshuffled = self.pixel_unshuffle(x)
+
+        # Convolution operation
+        x_conv = self.conv(x_unshuffled)
+
+        # Feature extraction with residual blocks
+        out = self.residual_blocks(x_conv)
+
+        # Reshape to restore original bf dimension
+        out = out.view(bs, f, out.size(1), out.size(2), out.size(3))
+
+        # Permute dimensions to reorder (if needed), e.g., swap channels and feature frames
+        out = out.permute(0, 2, 1, 3, 4)
+
+        return out
+
+
+class WanCamResidualBlock(nn.Module):
+    def __init__(self, dim, operation_settings={}):
+        super(WanCamResidualBlock, self).__init__()
+        self.conv1 = operation_settings.get("operations").Conv2d(dim, dim, kernel_size=3, padding=1, device=operation_settings.get("device"), dtype=operation_settings.get("dtype"))
+        self.relu = nn.ReLU(inplace=True)
+        self.conv2 = operation_settings.get("operations").Conv2d(dim, dim, kernel_size=3, padding=1, device=operation_settings.get("device"), dtype=operation_settings.get("dtype"))
+
+    def forward(self, x):
+        residual = x
+        out = self.relu(self.conv1(x))
+        out = self.conv2(out)
+        out += residual
+        return out
+
+
 class Head(nn.Module):
 
     def __init__(self, dim, out_dim, patch_size, eps=1e-6, operation_settings={}):
@@ -637,3 +691,92 @@ class VaceWanModel(WanModel):
         # unpatchify
         x = self.unpatchify(x, grid_sizes)
         return x
+
+class CameraWanModel(WanModel):
+    r"""
+    Wan diffusion backbone supporting both text-to-video and image-to-video.
+    """
+
+    def __init__(self,
+                 model_type='camera',
+                 patch_size=(1, 2, 2),
+                 text_len=512,
+                 in_dim=16,
+                 dim=2048,
+                 ffn_dim=8192,
+                 freq_dim=256,
+                 text_dim=4096,
+                 out_dim=16,
+                 num_heads=16,
+                 num_layers=32,
+                 window_size=(-1, -1),
+                 qk_norm=True,
+                 cross_attn_norm=True,
+                 eps=1e-6,
+                 flf_pos_embed_token_number=None,
+                 image_model=None,
+                 in_dim_control_adapter=24,
+                 device=None,
+                 dtype=None,
+                 operations=None,
+                 ):
+
+        super().__init__(model_type='i2v', patch_size=patch_size, text_len=text_len, in_dim=in_dim, dim=dim, ffn_dim=ffn_dim, freq_dim=freq_dim, text_dim=text_dim, out_dim=out_dim, num_heads=num_heads, num_layers=num_layers, window_size=window_size, qk_norm=qk_norm, cross_attn_norm=cross_attn_norm, eps=eps, flf_pos_embed_token_number=flf_pos_embed_token_number, image_model=image_model, device=device, dtype=dtype, operations=operations)
+        operation_settings = {"operations": operations, "device": device, "dtype": dtype}
+
+        self.control_adapter = WanCamAdapter(in_dim_control_adapter, dim, kernel_size=patch_size[1:], stride=patch_size[1:], operation_settings=operation_settings)
+
+
+    def forward_orig(
+        self,
+        x,
+        t,
+        context,
+        clip_fea=None,
+        freqs=None,
+        camera_conditions = None,
+        transformer_options={},
+        **kwargs,
+    ):
+        # embeddings
+        x = self.patch_embedding(x.float()).to(x.dtype)
+        if self.control_adapter is not None and camera_conditions is not None:
+            x_camera = self.control_adapter(camera_conditions).to(x.dtype)
+            x = x + x_camera
+        grid_sizes = x.shape[2:]
+        x = x.flatten(2).transpose(1, 2)
+
+        # time embeddings
+        e = self.time_embedding(
+            sinusoidal_embedding_1d(self.freq_dim, t).to(dtype=x[0].dtype))
+        e0 = self.time_projection(e).unflatten(1, (6, self.dim))
+
+        # context
+        context = self.text_embedding(context)
+
+        context_img_len = None
+        if clip_fea is not None:
+            if self.img_emb is not None:
+                context_clip = self.img_emb(clip_fea)  # bs x 257 x dim
+                context = torch.concat([context_clip, context], dim=1)
+            context_img_len = clip_fea.shape[-2]
+
+        patches_replace = transformer_options.get("patches_replace", {})
+        blocks_replace = patches_replace.get("dit", {})
+        for i, block in enumerate(self.blocks):
+            if ("double_block", i) in blocks_replace:
+                def block_wrap(args):
+                    out = {}
+                    out["img"] = block(args["img"], context=args["txt"], e=args["vec"], freqs=args["pe"], context_img_len=context_img_len)
+                    return out
+                out = blocks_replace[("double_block", i)]({"img": x, "txt": context, "vec": e0, "pe": freqs}, {"original_block": block_wrap})
+                x = out["img"]
+            else:
+                x = block(x, e=e0, freqs=freqs, context=context, context_img_len=context_img_len)
+
+        # head
+        x = self.head(x, e)
+
+        # unpatchify
+        x = self.unpatchify(x, grid_sizes)
+        return x

comfy/lora.py

@@ -286,6 +286,12 @@ def model_lora_keys_unet(model, key_map={}):
                     key_lora = k[len("diffusion_model."):-len(".weight")].replace(".", "_")
                     key_map["lycoris_{}".format(key_lora)] = k #SimpleTuner lycoris format
 
+    if isinstance(model, comfy.model_base.ACEStep):
+        for k in sdk:
+            if k.startswith("diffusion_model.") and k.endswith(".weight"): #Official ACE step lora format
+                key_lora = k[len("diffusion_model."):-len(".weight")]
+                key_map["{}".format(key_lora)] = k
+
     return key_map
 
 

comfy/model_base.py

@@ -924,6 +924,10 @@ class HunyuanVideo(BaseModel):
         if guiding_frame_index is not None:
             out['guiding_frame_index'] = comfy.conds.CONDRegular(torch.FloatTensor([guiding_frame_index]))
 
+        ref_latent = kwargs.get("ref_latent", None)
+        if ref_latent is not None:
+            out['ref_latent'] = comfy.conds.CONDRegular(self.process_latent_in(ref_latent))
+
         return out
 
     def scale_latent_inpaint(self, latent_image, **kwargs):
@@ -1075,6 +1079,17 @@ class WAN21_Vace(WAN21):
         out['vace_strength'] = comfy.conds.CONDConstant(vace_strength)
         return out
 
+class WAN21_Camera(WAN21):
+    def __init__(self, model_config, model_type=ModelType.FLOW, image_to_video=False, device=None):
+        super(WAN21, self).__init__(model_config, model_type, device=device, unet_model=comfy.ldm.wan.model.CameraWanModel)
+        self.image_to_video = image_to_video
+
+    def extra_conds(self, **kwargs):
+        out = super().extra_conds(**kwargs)
+        camera_conditions = kwargs.get("camera_conditions", None)
+        if camera_conditions is not None:
+            out['camera_conditions'] = comfy.conds.CONDRegular(camera_conditions)
+        return out
 
 class Hunyuan3Dv2(BaseModel):
     def __init__(self, model_config, model_type=ModelType.FLOW, device=None):

comfy/model_detection.py

@@ -361,6 +361,8 @@ def detect_unet_config(state_dict, key_prefix, metadata=None):
             dit_config["model_type"] = "vace"
             dit_config["vace_in_dim"] = state_dict['{}vace_patch_embedding.weight'.format(key_prefix)].shape[1]
             dit_config["vace_layers"] = count_blocks(state_dict_keys, '{}vace_blocks.'.format(key_prefix) + '{}.')
+        elif '{}control_adapter.conv.weight'.format(key_prefix) in state_dict_keys:
+            dit_config["model_type"] = "camera"
         else:
             if '{}img_emb.proj.0.bias'.format(key_prefix) in state_dict_keys:
                 dit_config["model_type"] = "i2v"

comfy/rmsnorm.py

@@ -30,7 +30,7 @@ if RMSNorm is None:
         def __init__(
             self,
             normalized_shape,
-            eps=None,
+            eps=1e-6,
             elementwise_affine=True,
             device=None,
             dtype=None,

comfy/supported_models.py

@@ -992,6 +992,16 @@ class WAN21_FunControl2V(WAN21_T2V):
         out = model_base.WAN21(self, image_to_video=False, device=device)
         return out
 
+class WAN21_Camera(WAN21_T2V):
+    unet_config = {
+        "image_model": "wan2.1",
+        "model_type": "camera",
+        "in_dim": 32,
+    }
+
+    def get_model(self, state_dict, prefix="", device=None):
+        out = model_base.WAN21_Camera(self, image_to_video=False, device=device)
+        return out
 class WAN21_Vace(WAN21_T2V):
     unet_config = {
         "image_model": "wan2.1",
@@ -1129,6 +1139,6 @@ class ACEStep(supported_models_base.BASE):
     def clip_target(self, state_dict={}):
         return supported_models_base.ClipTarget(comfy.text_encoders.ace.AceT5Tokenizer, comfy.text_encoders.ace.AceT5Model)
 
-models = [LotusD, Stable_Zero123, SD15_instructpix2pix, SD15, SD20, SD21UnclipL, SD21UnclipH, SDXL_instructpix2pix, SDXLRefiner, SDXL, SSD1B, KOALA_700M, KOALA_1B, Segmind_Vega, SD_X4Upscaler, Stable_Cascade_C, Stable_Cascade_B, SV3D_u, SV3D_p, SD3, StableAudio, AuraFlow, PixArtAlpha, PixArtSigma, HunyuanDiT, HunyuanDiT1, FluxInpaint, Flux, FluxSchnell, GenmoMochi, LTXV, HunyuanVideoSkyreelsI2V, HunyuanVideoI2V, HunyuanVideo, CosmosT2V, CosmosI2V, Lumina2, WAN21_T2V, WAN21_I2V, WAN21_FunControl2V, WAN21_Vace, Hunyuan3Dv2mini, Hunyuan3Dv2, HiDream, Chroma, ACEStep]
+models = [LotusD, Stable_Zero123, SD15_instructpix2pix, SD15, SD20, SD21UnclipL, SD21UnclipH, SDXL_instructpix2pix, SDXLRefiner, SDXL, SSD1B, KOALA_700M, KOALA_1B, Segmind_Vega, SD_X4Upscaler, Stable_Cascade_C, Stable_Cascade_B, SV3D_u, SV3D_p, SD3, StableAudio, AuraFlow, PixArtAlpha, PixArtSigma, HunyuanDiT, HunyuanDiT1, FluxInpaint, Flux, FluxSchnell, GenmoMochi, LTXV, HunyuanVideoSkyreelsI2V, HunyuanVideoI2V, HunyuanVideo, CosmosT2V, CosmosI2V, Lumina2, WAN21_T2V, WAN21_I2V, WAN21_FunControl2V, WAN21_Vace, WAN21_Camera, Hunyuan3Dv2mini, Hunyuan3Dv2, HiDream, Chroma, ACEStep]
 
 models += [SVD_img2vid]

comfy/utils.py

@@ -78,8 +78,6 @@ def load_torch_file(ckpt, safe_load=False, device=None, return_metadata=False):
             pl_sd = torch.load(ckpt, map_location=device, weights_only=True, **torch_args)
         else:
             pl_sd = torch.load(ckpt, map_location=device, pickle_module=comfy.checkpoint_pickle)
-        if "global_step" in pl_sd:
-            logging.debug(f"Global Step: {pl_sd['global_step']}")
         if "state_dict" in pl_sd:
             sd = pl_sd["state_dict"]
         else:

comfy_api/input/video_types.py

@@ -43,3 +43,13 @@ class VideoInput(ABC):
         components = self.get_components()
         return components.images.shape[2], components.images.shape[1]
 
+    def get_duration(self) -> float:
+        """
+        Returns the duration of the video in seconds.
+
+        Returns:
+            Duration in seconds
+        """
+        components = self.get_components()
+        frame_count = components.images.shape[0]
+        return float(frame_count / components.frame_rate)

comfy_api/input_impl/video_types.py

@@ -80,6 +80,38 @@ class VideoFromFile(VideoInput):
                     return stream.width, stream.height
         raise ValueError(f"No video stream found in file '{self.__file}'")
 
+    def get_duration(self) -> float:
+        """
+        Returns the duration of the video in seconds.
+
+        Returns:
+            Duration in seconds
+        """
+        if isinstance(self.__file, io.BytesIO):
+            self.__file.seek(0)
+        with av.open(self.__file, mode="r") as container:
+            if container.duration is not None:
+                return float(container.duration / av.time_base)
+
+            # Fallback: calculate from frame count and frame rate
+            video_stream = next(
+                (s for s in container.streams if s.type == "video"), None
+            )
+            if video_stream and video_stream.frames and video_stream.average_rate:
+                return float(video_stream.frames / video_stream.average_rate)
+
+            # Last resort: decode frames to count them
+            if video_stream and video_stream.average_rate:
+                frame_count = 0
+                container.seek(0)
+                for packet in container.demux(video_stream):
+                    for _ in packet.decode():
+                        frame_count += 1
+                if frame_count > 0:
+                    return float(frame_count / video_stream.average_rate)
+
+        raise ValueError(f"Could not determine duration for file '{self.__file}'")
+
     def get_components_internal(self, container: InputContainer) -> VideoComponents:
         # Get video frames
         frames = []

comfy_api_nodes/apinode_utils.py

@@ -1,7 +1,7 @@
 from __future__ import annotations
 import io
 import logging
-from typing import Optional
+from typing import Optional, Union
 from comfy.utils import common_upscale
 from comfy_api.input_impl import VideoFromFile
 from comfy_api.util import VideoContainer, VideoCodec
@@ -15,6 +15,7 @@ from comfy_api_nodes.apis.client import (
     UploadRequest,
     UploadResponse,
 )
+from server import PromptServer
 
 
 import numpy as np
@@ -60,7 +61,9 @@ def downscale_image_tensor(image, total_pixels=1536 * 1024) -> torch.Tensor:
     return s
 
 
-def validate_and_cast_response(response, timeout: int = None) -> torch.Tensor:
+def validate_and_cast_response(
+    response, timeout: int = None, node_id: Union[str, None] = None
+) -> torch.Tensor:
     """Validates and casts a response to a torch.Tensor.
 
     Args:
@@ -94,6 +97,10 @@ def validate_and_cast_response(response, timeout: int = None) -> torch.Tensor:
             img = Image.open(io.BytesIO(img_data))
 
         elif image_url:
+            if node_id:
+                PromptServer.instance.send_progress_text(
+                    f"Result URL: {image_url}", node_id
+                )
             img_response = requests.get(image_url, timeout=timeout)
             if img_response.status_code != 200:
                 raise ValueError("Failed to download the image")

comfy_api_nodes/apis/client.py

@@ -94,15 +94,19 @@ from __future__ import annotations
 import logging
 import time
 import io
-from typing import Dict, Type, Optional, Any, TypeVar, Generic, Callable
+import socket
+from typing import Dict, Type, Optional, Any, TypeVar, Generic, Callable, Tuple
 from enum import Enum
 import json
 import requests
-from urllib.parse import urljoin
+from urllib.parse import urljoin, urlparse
 from pydantic import BaseModel, Field
+import uuid # For generating unique operation IDs
 
+from server import PromptServer
 from comfy.cli_args import args
 from comfy import utils
+from . import request_logger
 
 T = TypeVar("T", bound=BaseModel)
 R = TypeVar("R", bound=BaseModel)
@@ -111,6 +115,21 @@ P = TypeVar("P", bound=BaseModel)  # For poll response
 PROGRESS_BAR_MAX = 100
 
 
+class NetworkError(Exception):
+    """Base exception for network-related errors with diagnostic information."""
+    pass
+
+
+class LocalNetworkError(NetworkError):
+    """Exception raised when local network connectivity issues are detected."""
+    pass
+
+
+class ApiServerError(NetworkError):
+    """Exception raised when the API server is unreachable but internet is working."""
+    pass
+
+
 class EmptyRequest(BaseModel):
     """Base class for empty request bodies.
     For GET requests, fields will be sent as query parameters."""
@@ -141,7 +160,7 @@ class HttpMethod(str, Enum):
 
 class ApiClient:
     """
-    Client for making HTTP requests to an API with authentication and error handling.
+    Client for making HTTP requests to an API with authentication, error handling, and retry logic.
     """
 
     def __init__(
@@ -151,12 +170,26 @@ class ApiClient:
         comfy_api_key: Optional[str] = None,
         timeout: float = 3600.0,
         verify_ssl: bool = True,
+        max_retries: int = 3,
+        retry_delay: float = 1.0,
+        retry_backoff_factor: float = 2.0,
+        retry_status_codes: Optional[Tuple[int, ...]] = None,
     ):
         self.base_url = base_url
         self.auth_token = auth_token
         self.comfy_api_key = comfy_api_key
         self.timeout = timeout
         self.verify_ssl = verify_ssl
+        self.max_retries = max_retries
+        self.retry_delay = retry_delay
+        self.retry_backoff_factor = retry_backoff_factor
+        # Default retry status codes: 408 (Request Timeout), 429 (Too Many Requests),
+        # 500, 502, 503, 504 (Server Errors)
+        self.retry_status_codes = retry_status_codes or (408, 429, 500, 502, 503, 504)
+
+    def _generate_operation_id(self, path: str) -> str:
+        """Generates a unique operation ID for logging."""
+        return f"{path.strip('/').replace('/', '_')}_{uuid.uuid4().hex[:8]}"
 
     def _create_json_payload_args(
         self,
@@ -211,6 +244,56 @@ class ApiClient:
 
         return headers
 
+    def _check_connectivity(self, target_url: str) -> Dict[str, bool]:
+        """
+        Check connectivity to determine if network issues are local or server-related.
+
+        Args:
+            target_url: URL to check connectivity to
+
+        Returns:
+            Dictionary with connectivity status details
+        """
+        results = {
+            "internet_accessible": False,
+            "api_accessible": False,
+            "is_local_issue": False,
+            "is_api_issue": False
+        }
+
+        # First check basic internet connectivity using a reliable external site
+        try:
+            # Use a reliable external domain for checking basic connectivity
+            check_response = requests.get("https://www.google.com",
+                                         timeout=5.0,
+                                         verify=self.verify_ssl)
+            if check_response.status_code < 500:
+                results["internet_accessible"] = True
+        except (requests.RequestException, socket.error):
+            results["internet_accessible"] = False
+            results["is_local_issue"] = True
+            return results
+
+        # Now check API server connectivity
+        try:
+            # Extract domain from the target URL to do a simpler health check
+            parsed_url = urlparse(target_url)
+            api_base = f"{parsed_url.scheme}://{parsed_url.netloc}"
+
+            # Try to reach the API domain
+            api_response = requests.get(f"{api_base}/health", timeout=5.0, verify=self.verify_ssl)
+            if api_response.status_code < 500:
+                results["api_accessible"] = True
+            else:
+                results["api_accessible"] = False
+                results["is_api_issue"] = True
+        except requests.RequestException:
+            results["api_accessible"] = False
+            # If we can reach the internet but not the API, it's an API issue
+            results["is_api_issue"] = True
+
+        return results
+
     def request(
         self,
         method: str,
@@ -221,9 +304,10 @@ class ApiClient:
         headers: Optional[Dict[str, str]] = None,
         content_type: str = "application/json",
         multipart_parser: Callable = None,
+        retry_count: int = 0,  # Used internally for tracking retries
     ) -> Dict[str, Any]:
         """
-        Make an HTTP request to the API
+        Make an HTTP request to the API with automatic retries for transient errors.
 
         Args:
             method: HTTP method (GET, POST, etc.)
@@ -233,12 +317,15 @@ class ApiClient:
             files: Files to upload
             headers: Additional headers
             content_type: Content type of the request. Defaults to application/json.
+            retry_count: Internal parameter for tracking retries, do not set manually
 
         Returns:
             Parsed JSON response
 
         Raises:
-            requests.RequestException: If the request fails
+            LocalNetworkError: If local network connectivity issues are detected
+            ApiServerError: If the API server is unreachable but internet is working
+            Exception: For other request failures
         """
         url = urljoin(self.base_url, path)
         self.check_auth(self.auth_token, self.comfy_api_key)
@@ -265,6 +352,16 @@ class ApiClient:
         else:
             payload_args = self._create_json_payload_args(data, request_headers)
 
+        operation_id = self._generate_operation_id(path)
+        request_logger.log_request_response(
+            operation_id=operation_id,
+            request_method=method,
+            request_url=url,
+            request_headers=request_headers,
+            request_params=params,
+            request_data=data if content_type == "application/json" else "[form-data or other]"
+        )
+
         try:
             response = requests.request(
                 method=method,
@@ -275,50 +372,228 @@ class ApiClient:
                 **payload_args,
             )
 
+            # Check if we should retry based on status code
+            if (response.status_code in self.retry_status_codes and
+                retry_count < self.max_retries):
+
+                # Calculate delay with exponential backoff
+                delay = self.retry_delay * (self.retry_backoff_factor ** retry_count)
+
+                logging.warning(
+                    f"Request failed with status {response.status_code}. "
+                    f"Retrying in {delay:.2f}s ({retry_count + 1}/{self.max_retries})"
+                )
+
+                time.sleep(delay)
+                return self.request(
+                    method=method,
+                    path=path,
+                    params=params,
+                    data=data,
+                    files=files,
+                    headers=headers,
+                    content_type=content_type,
+                    multipart_parser=multipart_parser,
+                    retry_count=retry_count + 1,
+                )
+
             # Raise exception for error status codes
             response.raise_for_status()
-        except requests.ConnectionError:
-            raise Exception(
-                f"Unable to connect to the API server at {self.base_url}. Please check your internet connection or verify the service is available."
+
+            # Log successful response
+            response_content_to_log = response.content
+            try:
+                # Attempt to parse JSON for prettier logging, fallback to raw content
+                response_content_to_log = response.json()
+            except json.JSONDecodeError:
+                pass # Keep as bytes/str if not JSON
+
+            request_logger.log_request_response(
+                operation_id=operation_id,
+                request_method=method, # Pass request details again for context in log
+                request_url=url,
+                response_status_code=response.status_code,
+                response_headers=dict(response.headers),
+                response_content=response_content_to_log
             )
 
-        except requests.Timeout:
-            raise Exception(
-                f"Request timed out after {self.timeout} seconds. The server might be experiencing high load or the operation is taking longer than expected."
+        except requests.ConnectionError as e:
+            error_message = f"ConnectionError: {str(e)}"
+            request_logger.log_request_response(
+                operation_id=operation_id,
+                request_method=method,
+                request_url=url,
+                error_message=error_message
             )
+            # Only perform connectivity check if we've exhausted all retries
+            if retry_count >= self.max_retries:
+                # Check connectivity to determine if it's a local or API issue
+                connectivity = self._check_connectivity(self.base_url)
+
+                if connectivity["is_local_issue"]:
+                    raise LocalNetworkError(
+                        "Unable to connect to the API server due to local network issues. "
+                        "Please check your internet connection and try again."
+                    ) from e
+                elif connectivity["is_api_issue"]:
+                    raise ApiServerError(
+                        f"The API server at {self.base_url} is currently unreachable. "
+                        f"The service may be experiencing issues. Please try again later."
+                    ) from e
+
+            # If we haven't exhausted retries yet, retry the request
+            if retry_count < self.max_retries:
+                delay = self.retry_delay * (self.retry_backoff_factor ** retry_count)
+                logging.warning(
+                    f"Connection error: {str(e)}. "
+                    f"Retrying in {delay:.2f}s ({retry_count + 1}/{self.max_retries})"
+                )
+                time.sleep(delay)
+                return self.request(
+                    method=method,
+                    path=path,
+                    params=params,
+                    data=data,
+                    files=files,
+                    headers=headers,
+                    content_type=content_type,
+                    multipart_parser=multipart_parser,
+                    retry_count=retry_count + 1,
+                )
+
+            # If we've exhausted retries and didn't identify the specific issue,
+            # raise a generic exception
+            final_error_message = (
+                f"Unable to connect to the API server after {self.max_retries} attempts. "
+                f"Please check your internet connection or try again later."
+            )
+            request_logger.log_request_response( # Log final failure
+                operation_id=operation_id,
+                request_method=method, request_url=url,
+                error_message=final_error_message
+            )
+            raise Exception(final_error_message) from e
+
+        except requests.Timeout as e:
+            error_message = f"Timeout: {str(e)}"
+            request_logger.log_request_response(
+                operation_id=operation_id,
+                request_method=method, request_url=url,
+                error_message=error_message
+            )
+            # Retry timeouts if we haven't exhausted retries
+            if retry_count < self.max_retries:
+                delay = self.retry_delay * (self.retry_backoff_factor ** retry_count)
+                logging.warning(
+                    f"Request timed out. "
+                    f"Retrying in {delay:.2f}s ({retry_count + 1}/{self.max_retries})"
+                )
+                time.sleep(delay)
+                return self.request(
+                    method=method,
+                    path=path,
+                    params=params,
+                    data=data,
+                    files=files,
+                    headers=headers,
+                    content_type=content_type,
+                    multipart_parser=multipart_parser,
+                    retry_count=retry_count + 1,
+                )
+            final_error_message = (
+                f"Request timed out after {self.timeout} seconds and {self.max_retries} retry attempts. "
+                f"The server might be experiencing high load or the operation is taking longer than expected."
+            )
+            request_logger.log_request_response( # Log final failure
+                operation_id=operation_id,
+                request_method=method, request_url=url,
+                error_message=final_error_message
+            )
+            raise Exception(final_error_message) from e
 
         except requests.HTTPError as e:
             status_code = e.response.status_code if hasattr(e, "response") else None
-            error_message = f"HTTP Error: {str(e)}"
+            original_error_message = f"HTTP Error: {str(e)}"
+            error_content_for_log = None
+            if hasattr(e, "response") and e.response is not None:
+                error_content_for_log = e.response.content
+                try:
+                    error_content_for_log = e.response.json()
+                except json.JSONDecodeError:
+                    pass
+
+
+            # Try to extract detailed error message from JSON response for user display
+            # but log the full error content.
+            user_display_error_message = original_error_message
 
-            # Try to extract detailed error message from JSON response
             try:
-                if hasattr(e, "response") and e.response.content:
+                if hasattr(e, "response") and e.response is not None and e.response.content:
                     error_json = e.response.json()
                     if "error" in error_json and "message" in error_json["error"]:
-                        error_message = f"API Error: {error_json['error']['message']}"
+                        user_display_error_message = f"API Error: {error_json['error']['message']}"
                         if "type" in error_json["error"]:
-                            error_message += f" (Type: {error_json['error']['type']})"
+                            user_display_error_message += f" (Type: {error_json['error']['type']})"
+                    elif isinstance(error_json, dict): # Handle cases where error is just a JSON dict
+                        user_display_error_message = f"API Error: {json.dumps(error_json)}"
+                    else: # Non-dict JSON error
+                        user_display_error_message = f"API Error: {str(error_json)}"
+            except json.JSONDecodeError:
+                # If not JSON, use the raw content if it's not too long, or a summary
+                if hasattr(e, "response") and e.response is not None and e.response.content:
+                    raw_content = e.response.content.decode(errors='ignore')
+                    if len(raw_content) < 200: # Arbitrary limit for display
+                        user_display_error_message = f"API Error (raw): {raw_content}"
                     else:
-                        error_message = f"API Error: {error_json}"
-            except Exception as json_error:
-                # If we can't parse the JSON, fall back to the original error message
-                logging.debug(
-                    f"[DEBUG] Failed to parse error response: {str(json_error)}"
+                        user_display_error_message = f"API Error (raw, status {status_code})"
+
+            request_logger.log_request_response(
+                operation_id=operation_id,
+                request_method=method, request_url=url,
+                response_status_code=status_code,
+                response_headers=dict(e.response.headers) if hasattr(e, "response") and e.response is not None else None,
+                response_content=error_content_for_log,
+                error_message=original_error_message # Log the original exception string as error
+            )
+
+            logging.debug(f"[DEBUG] API Error: {user_display_error_message} (Status: {status_code})")
+            if hasattr(e, "response") and e.response is not None and e.response.content:
+                logging.debug(f"[DEBUG] Response content: {e.response.content}")
+
+            # Retry if the status code is in our retry list and we haven't exhausted retries
+            if (status_code in self.retry_status_codes and
+                retry_count < self.max_retries):
+
+                delay = self.retry_delay * (self.retry_backoff_factor ** retry_count)
+                logging.warning(
+                    f"HTTP error {status_code}. "
+                    f"Retrying in {delay:.2f}s ({retry_count + 1}/{self.max_retries})"
+                )
+                time.sleep(delay)
+                return self.request(
+                    method=method,
+                    path=path,
+                    params=params,
+                    data=data,
+                    files=files,
+                    headers=headers,
+                    content_type=content_type,
+                    multipart_parser=multipart_parser,
+                    retry_count=retry_count + 1,
                 )
 
-            logging.debug(f"[DEBUG] API Error: {error_message} (Status: {status_code})")
-            if hasattr(e, "response") and e.response.content:
-                logging.debug(f"[DEBUG] Response content: {e.response.content}")
+            # Specific error messages for common status codes for user display
             if status_code == 401:
-                error_message = "Unauthorized: Please login first to use this node."
-            if status_code == 402:
-                error_message = "Payment Required: Please add credits to your account to use this node."
-            if status_code == 409:
-                error_message = "There is a problem with your account. Please contact support@comfy.org. "
-            if status_code == 429:
-                error_message = "Rate Limit Exceeded: Please try again later."
-            raise Exception(error_message)
+                user_display_error_message = "Unauthorized: Please login first to use this node."
+            elif status_code == 402:
+                user_display_error_message = "Payment Required: Please add credits to your account to use this node."
+            elif status_code == 409:
+                user_display_error_message = "There is a problem with your account. Please contact support@comfy.org."
+            elif status_code == 429:
+                user_display_error_message = "Rate Limit Exceeded: Please try again later."
+            # else, user_display_error_message remains as parsed from response or original HTTPError string
+
+            raise Exception(user_display_error_message) # Raise with the user-friendly message
 
         # Parse and return JSON response
         if response.content:
@@ -336,26 +611,126 @@ class ApiClient:
         upload_url: str,
         file: io.BytesIO | str,
         content_type: str | None = None,
+        max_retries: int = 3,
+        retry_delay: float = 1.0,
+        retry_backoff_factor: float = 2.0,
     ):
-        """Upload a file to the API. Make sure the file has a filename equal to what the url expects.
+        """Upload a file to the API with retry logic.
 
         Args:
             upload_url: The URL to upload to
             file: Either a file path string, BytesIO object, or tuple of (file_path, filename)
-            mime_type: Optional mime type to set for the upload
+            content_type: Optional mime type to set for the upload
+            max_retries: Maximum number of retry attempts
+            retry_delay: Initial delay between retries in seconds
+            retry_backoff_factor: Multiplier for the delay after each retry
         """
         headers = {}
         if content_type:
             headers["Content-Type"] = content_type
 
+        # Prepare the file data
         if isinstance(file, io.BytesIO):
             file.seek(0)  # Ensure we're at the start of the file
             data = file.read()
-            return requests.put(upload_url, data=data, headers=headers)
         elif isinstance(file, str):
             with open(file, "rb") as f:
                 data = f.read()
-                return requests.put(upload_url, data=data, headers=headers)
+        else:
+            raise ValueError("File must be either a BytesIO object or a file path string")
+
+        # Try the upload with retries
+        last_exception = None
+        operation_id = f"upload_{upload_url.split('/')[-1]}_{uuid.uuid4().hex[:8]}" # Simplified ID for uploads
+
+        # Log initial attempt (without full file data for brevity)
+        request_logger.log_request_response(
+            operation_id=operation_id,
+            request_method="PUT",
+            request_url=upload_url,
+            request_headers=headers,
+            request_data=f"[File data of type {content_type or 'unknown'}, size {len(data)} bytes]"
+        )
+
+        for retry_attempt in range(max_retries + 1):
+            try:
+                response = requests.put(upload_url, data=data, headers=headers)
+                response.raise_for_status()
+                request_logger.log_request_response(
+                    operation_id=operation_id,
+                    request_method="PUT", request_url=upload_url, # For context
+                    response_status_code=response.status_code,
+                    response_headers=dict(response.headers),
+                    response_content="File uploaded successfully." # Or response.text if available
+                )
+                return response
+
+            except (requests.ConnectionError, requests.Timeout, requests.HTTPError) as e:
+                last_exception = e
+                error_message_for_log = f"{type(e).__name__}: {str(e)}"
+                response_content_for_log = None
+                status_code_for_log = None
+                headers_for_log = None
+
+                if hasattr(e, 'response') and e.response is not None:
+                    status_code_for_log = e.response.status_code
+                    headers_for_log = dict(e.response.headers)
+                    try:
+                        response_content_for_log = e.response.json()
+                    except json.JSONDecodeError:
+                        response_content_for_log = e.response.content
+
+
+                request_logger.log_request_response(
+                    operation_id=operation_id,
+                    request_method="PUT", request_url=upload_url,
+                    response_status_code=status_code_for_log,
+                    response_headers=headers_for_log,
+                    response_content=response_content_for_log,
+                    error_message=error_message_for_log
+                )
+
+                if retry_attempt < max_retries:
+                    delay = retry_delay * (retry_backoff_factor ** retry_attempt)
+                    logging.warning(
+                        f"File upload failed: {str(e)}. "
+                        f"Retrying in {delay:.2f}s ({retry_attempt + 1}/{max_retries})"
+                    )
+                    time.sleep(delay)
+                else:
+                    break # Max retries reached
+
+        # If we've exhausted all retries, determine the final error type and raise
+        final_error_message = f"Failed to upload file after {max_retries + 1} attempts. Error: {str(last_exception)}"
+        try:
+            # Check basic internet connectivity
+            check_response = requests.get("https://www.google.com", timeout=5.0, verify=True) # Assuming verify=True is desired
+            if check_response.status_code >= 500: # Google itself has an issue (rare)
+                 final_error_message = (f"Failed to upload file. Internet connectivity check to Google failed "
+                                       f"(status {check_response.status_code}). Original error: {str(last_exception)}")
+                 # Not raising LocalNetworkError here as Google itself might be down.
+            # If Google is reachable, the issue is likely with the upload server or a more specific local problem
+            # not caught by a simple Google ping (e.g., DNS for the specific upload URL, firewall).
+            # The original last_exception is probably most relevant.
+
+        except (requests.RequestException, socket.error) as conn_check_exc:
+            # Could not reach Google, likely a local network issue
+            final_error_message = (f"Failed to upload file due to network connectivity issues "
+                                   f"(cannot reach Google: {str(conn_check_exc)}). "
+                                   f"Original upload error: {str(last_exception)}")
+            request_logger.log_request_response( # Log final failure reason
+                operation_id=operation_id,
+                request_method="PUT", request_url=upload_url,
+                error_message=final_error_message
+            )
+            raise LocalNetworkError(final_error_message) from last_exception
+
+        request_logger.log_request_response( # Log final failure reason if not LocalNetworkError
+            operation_id=operation_id,
+            request_method="PUT", request_url=upload_url,
+            error_message=final_error_message
+        )
+        raise Exception(final_error_message) from last_exception
 
 
 class ApiEndpoint(Generic[T, R]):
@@ -403,6 +778,9 @@ class SynchronousOperation(Generic[T, R]):
         verify_ssl: bool = True,
         content_type: str = "application/json",
         multipart_parser: Callable = None,
+        max_retries: int = 3,
+        retry_delay: float = 1.0,
+        retry_backoff_factor: float = 2.0,
     ):
         self.endpoint = endpoint
         self.request = request
@@ -419,8 +797,12 @@ class SynchronousOperation(Generic[T, R]):
         self.files = files
         self.content_type = content_type
         self.multipart_parser = multipart_parser
+        self.max_retries = max_retries
+        self.retry_delay = retry_delay
+        self.retry_backoff_factor = retry_backoff_factor
+
     def execute(self, client: Optional[ApiClient] = None) -> R:
-        """Execute the API operation using the provided client or create one"""
+        """Execute the API operation using the provided client or create one with retry support"""
         try:
             # Create client if not provided
             if client is None:
@@ -430,6 +812,9 @@ class SynchronousOperation(Generic[T, R]):
                     comfy_api_key=self.comfy_api_key,
                     timeout=self.timeout,
                     verify_ssl=self.verify_ssl,
+                    max_retries=self.max_retries,
+                    retry_delay=self.retry_delay,
+                    retry_backoff_factor=self.retry_backoff_factor,
                 )
 
             # Convert request model to dict, but use None for EmptyRequest
@@ -443,11 +828,6 @@ class SynchronousOperation(Generic[T, R]):
                     if isinstance(value, Enum):
                         request_dict[key] = value.value
 
-            if request_dict:
-               for key, value in request_dict.items():
-                   if isinstance(value, Enum):
-                       request_dict[key] = value.value
-
             # Debug log for request
             logging.debug(
                 f"[DEBUG] API Request: {self.endpoint.method.value} {self.endpoint.path}"
@@ -455,7 +835,7 @@ class SynchronousOperation(Generic[T, R]):
             logging.debug(f"[DEBUG] Request Data: {json.dumps(request_dict, indent=2)}")
             logging.debug(f"[DEBUG] Query Params: {self.endpoint.query_params}")
 
-            # Make the request
+            # Make the request with built-in retry
             resp = client.request(
                 method=self.endpoint.method.value,
                 path=self.endpoint.path,
@@ -476,8 +856,18 @@ class SynchronousOperation(Generic[T, R]):
             # Parse and return the response
             return self._parse_response(resp)
 
+        except LocalNetworkError as e:
+            # Propagate specific network error types
+            logging.error(f"[ERROR] Local network error: {str(e)}")
+            raise
+
+        except ApiServerError as e:
+            # Propagate API server errors
+            logging.error(f"[ERROR] API server error: {str(e)}")
+            raise
+
         except Exception as e:
-            logging.error(f"[DEBUG] API Exception: {str(e)}")
+            logging.error(f"[ERROR] API Exception: {str(e)}")
             raise Exception(str(e))
 
     def _parse_response(self, resp):
@@ -511,12 +901,19 @@ class PollingOperation(Generic[T, R]):
         failed_statuses: list,
         status_extractor: Callable[[R], str],
         progress_extractor: Callable[[R], float] = None,
+        result_url_extractor: Callable[[R], str] = None,
         request: Optional[T] = None,
         api_base: str | None = None,
         auth_token: Optional[str] = None,
         comfy_api_key: Optional[str] = None,
         auth_kwargs: Optional[Dict[str,str]] = None,
         poll_interval: float = 5.0,
+        max_poll_attempts: int = 120,  # Default max polling attempts (10 minutes with 5s interval)
+        max_retries: int = 3,  # Max retries per individual API call
+        retry_delay: float = 1.0,
+        retry_backoff_factor: float = 2.0,
+        estimated_duration: Optional[float] = None,
+        node_id: Optional[str] = None,
     ):
         self.poll_endpoint = poll_endpoint
         self.request = request
@@ -527,12 +924,19 @@ class PollingOperation(Generic[T, R]):
             self.auth_token = auth_kwargs.get("auth_token", self.auth_token)
             self.comfy_api_key = auth_kwargs.get("comfy_api_key", self.comfy_api_key)
         self.poll_interval = poll_interval
+        self.max_poll_attempts = max_poll_attempts
+        self.max_retries = max_retries
+        self.retry_delay = retry_delay
+        self.retry_backoff_factor = retry_backoff_factor
+        self.estimated_duration = estimated_duration
 
         # Polling configuration
         self.status_extractor = status_extractor or (
             lambda x: getattr(x, "status", None)
         )
         self.progress_extractor = progress_extractor
+        self.result_url_extractor = result_url_extractor
+        self.node_id = node_id
         self.completed_statuses = completed_statuses
         self.failed_statuses = failed_statuses
 
@@ -548,11 +952,46 @@ class PollingOperation(Generic[T, R]):
                     base_url=self.api_base,
                     auth_token=self.auth_token,
                     comfy_api_key=self.comfy_api_key,
+                    max_retries=self.max_retries,
+                    retry_delay=self.retry_delay,
+                    retry_backoff_factor=self.retry_backoff_factor,
                 )
             return self._poll_until_complete(client)
+        except LocalNetworkError as e:
+            # Provide clear message for local network issues
+            raise Exception(
+                f"Polling failed due to local network issues. Please check your internet connection. "
+                f"Details: {str(e)}"
+            ) from e
+        except ApiServerError as e:
+            # Provide clear message for API server issues
+            raise Exception(
+                f"Polling failed due to API server issues. The service may be experiencing problems. "
+                f"Please try again later. Details: {str(e)}"
+            ) from e
         except Exception as e:
             raise Exception(f"Error during polling: {str(e)}")
 
+    def _display_text_on_node(self, text: str):
+        """Sends text to the client which will be displayed on the node in the UI"""
+        if not self.node_id:
+            return
+
+        PromptServer.instance.send_progress_text(text, self.node_id)
+
+    def _display_time_progress_on_node(self, time_completed: int):
+        if not self.node_id:
+            return
+
+        if self.estimated_duration is not None:
+            estimated_time_remaining = max(
+                0, int(self.estimated_duration) - int(time_completed)
+            )
+            message = f"Task in progress: {time_completed:.0f}s (~{estimated_time_remaining:.0f}s remaining)"
+        else:
+            message = f"Task in progress: {time_completed:.0f}s"
+        self._display_text_on_node(message)
+
     def _check_task_status(self, response: R) -> TaskStatus:
         """Check task status using the status extractor function"""
         try:
@@ -569,10 +1008,13 @@ class PollingOperation(Generic[T, R]):
     def _poll_until_complete(self, client: ApiClient) -> R:
         """Poll until the task is complete"""
         poll_count = 0
+        consecutive_errors = 0
+        max_consecutive_errors = min(5, self.max_retries * 2)  # Limit consecutive errors
+
         if self.progress_extractor:
             progress = utils.ProgressBar(PROGRESS_BAR_MAX)
 
-        while True:
+        while poll_count < self.max_poll_attempts:
             try:
                 poll_count += 1
                 logging.debug(f"[DEBUG] Polling attempt #{poll_count}")
@@ -599,8 +1041,12 @@ class PollingOperation(Generic[T, R]):
                     data=request_dict,
                 )
 
+                # Successfully got a response, reset consecutive error count
+                consecutive_errors = 0
+
                 # Parse response
                 response_obj = self.poll_endpoint.response_model.model_validate(resp)
+
                 # Check if task is complete
                 status = self._check_task_status(response_obj)
                 logging.debug(f"[DEBUG] Task Status: {status}")
@@ -612,7 +1058,15 @@ class PollingOperation(Generic[T, R]):
                         progress.update_absolute(new_progress, total=PROGRESS_BAR_MAX)
 
                 if status == TaskStatus.COMPLETED:
-                    logging.debug("[DEBUG] Task completed successfully")
+                    message = "Task completed successfully"
+                    if self.result_url_extractor:
+                        result_url = self.result_url_extractor(response_obj)
+                        if result_url:
+                            message = f"Result URL: {result_url}"
+                    else:
+                        message = "Task completed successfully!"
+                    logging.debug(f"[DEBUG] {message}")
+                    self._display_text_on_node(message)
                     self.final_response = response_obj
                     if self.progress_extractor:
                         progress.update(100)
@@ -628,8 +1082,43 @@ class PollingOperation(Generic[T, R]):
                 logging.debug(
                     f"[DEBUG] Waiting {self.poll_interval} seconds before next poll"
                 )
+                for i in range(int(self.poll_interval)):
+                    time_completed = (poll_count * self.poll_interval) + i
+                    self._display_time_progress_on_node(time_completed)
+                    time.sleep(1)
+
+            except (LocalNetworkError, ApiServerError) as e:
+                # For network-related errors, increment error count and potentially abort
+                consecutive_errors += 1
+                if consecutive_errors >= max_consecutive_errors:
+                    raise Exception(
+                        f"Polling aborted after {consecutive_errors} consecutive network errors: {str(e)}"
+                    ) from e
+
+                # Log the error but continue polling
+                logging.warning(
+                    f"Network error during polling (attempt {poll_count}/{self.max_poll_attempts}): {str(e)}. "
+                    f"Will retry in {self.poll_interval} seconds."
+                )
                 time.sleep(self.poll_interval)
 
             except Exception as e:
+                # For other errors, increment count and potentially abort
+                consecutive_errors += 1
+                if consecutive_errors >= max_consecutive_errors or status == TaskStatus.FAILED:
+                    raise Exception(
+                        f"Polling aborted after {consecutive_errors} consecutive errors: {str(e)}"
+                    ) from e
+
                 logging.error(f"[DEBUG] Polling error: {str(e)}")
-                raise Exception(f"Error while polling: {str(e)}")
+                logging.warning(
+                    f"Error during polling (attempt {poll_count}/{self.max_poll_attempts}): {str(e)}. "
+                    f"Will retry in {self.poll_interval} seconds."
+                )
+                time.sleep(self.poll_interval)
+
+        # If we've exhausted all polling attempts
+        raise Exception(
+            f"Polling timed out after {poll_count} attempts ({poll_count * self.poll_interval} seconds). "
+            f"The operation may still be running on the server but is taking longer than expected."
+        )

comfy_api_nodes/apis/request_logger.py

@@ -0,0 +1,125 @@
+import os
+import datetime
+import json
+import logging
+import folder_paths
+
+# Get the logger instance
+logger = logging.getLogger(__name__)
+
+def get_log_directory():
+    """
+    Ensures the API log directory exists within ComfyUI's temp directory
+    and returns its path.
+    """
+    base_temp_dir = folder_paths.get_temp_directory()
+    log_dir = os.path.join(base_temp_dir, "api_logs")
+    try:
+        os.makedirs(log_dir, exist_ok=True)
+    except Exception as e:
+        logger.error(f"Error creating API log directory {log_dir}: {e}")
+        # Fallback to base temp directory if sub-directory creation fails
+        return base_temp_dir
+    return log_dir
+
+def _format_data_for_logging(data):
+    """Helper to format data (dict, str, bytes) for logging."""
+    if isinstance(data, bytes):
+        try:
+            return data.decode('utf-8')  # Try to decode as text
+        except UnicodeDecodeError:
+            return f"[Binary data of length {len(data)} bytes]"
+    elif isinstance(data, (dict, list)):
+        try:
+            return json.dumps(data, indent=2, ensure_ascii=False)
+        except TypeError:
+            return str(data) # Fallback for non-serializable objects
+    return str(data)
+
+def log_request_response(
+    operation_id: str,
+    request_method: str,
+    request_url: str,
+    request_headers: dict | None = None,
+    request_params: dict | None = None,
+    request_data: any = None,
+    response_status_code: int | None = None,
+    response_headers: dict | None = None,
+    response_content: any = None,
+    error_message: str | None = None
+):
+    """
+    Logs API request and response details to a file in the temp/api_logs directory.
+    """
+    log_dir = get_log_directory()
+    timestamp = datetime.datetime.now().strftime("%Y%m%d_%H%M%S_%f")
+    filename = f"{timestamp}_{operation_id.replace('/', '_').replace(':', '_')}.log"
+    filepath = os.path.join(log_dir, filename)
+
+    log_content = []
+
+    log_content.append(f"Timestamp: {datetime.datetime.now().isoformat()}")
+    log_content.append(f"Operation ID: {operation_id}")
+    log_content.append("-" * 30 + " REQUEST " + "-" * 30)
+    log_content.append(f"Method: {request_method}")
+    log_content.append(f"URL: {request_url}")
+    if request_headers:
+        log_content.append(f"Headers:\n{_format_data_for_logging(request_headers)}")
+    if request_params:
+        log_content.append(f"Params:\n{_format_data_for_logging(request_params)}")
+    if request_data:
+        log_content.append(f"Data/Body:\n{_format_data_for_logging(request_data)}")
+
+    log_content.append("\n" + "-" * 30 + " RESPONSE " + "-" * 30)
+    if response_status_code is not None:
+        log_content.append(f"Status Code: {response_status_code}")
+    if response_headers:
+        log_content.append(f"Headers:\n{_format_data_for_logging(response_headers)}")
+    if response_content:
+        log_content.append(f"Content:\n{_format_data_for_logging(response_content)}")
+    if error_message:
+        log_content.append(f"Error:\n{error_message}")
+
+    try:
+        with open(filepath, "w", encoding="utf-8") as f:
+            f.write("\n".join(log_content))
+        logger.debug(f"API log saved to: {filepath}")
+    except Exception as e:
+        logger.error(f"Error writing API log to {filepath}: {e}")
+
+if __name__ == '__main__':
+    # Example usage (for testing the logger directly)
+    logger.setLevel(logging.DEBUG)
+    # Mock folder_paths for direct execution if not running within ComfyUI full context
+    if not hasattr(folder_paths, 'get_temp_directory'):
+        class MockFolderPaths:
+            def get_temp_directory(self):
+                # Create a local temp dir for testing if needed
+                p = os.path.join(os.path.dirname(__file__), 'temp_test_logs')
+                os.makedirs(p, exist_ok=True)
+                return p
+        folder_paths = MockFolderPaths()
+
+    log_request_response(
+        operation_id="test_operation_get",
+        request_method="GET",
+        request_url="https://api.example.com/test",
+        request_headers={"Authorization": "Bearer testtoken"},
+        request_params={"param1": "value1"},
+        response_status_code=200,
+        response_content={"message": "Success!"}
+    )
+    log_request_response(
+        operation_id="test_operation_post_error",
+        request_method="POST",
+        request_url="https://api.example.com/submit",
+        request_data={"key": "value", "nested": {"num": 123}},
+        error_message="Connection timed out"
+    )
+    log_request_response(
+        operation_id="test_binary_response",
+        request_method="GET",
+        request_url="https://api.example.com/image.png",
+        response_status_code=200,
+        response_content=b'\x89PNG\r\n\x1a\n\x00\x00\x00\rIHDR...' # Sample binary data
+    )

comfy_api_nodes/nodes_bfl.py

@@ -1,5 +1,6 @@
 import io
 from inspect import cleandoc
+from typing import Union
 from comfy.comfy_types.node_typing import IO, ComfyNodeABC
 from comfy_api_nodes.apis.bfl_api import (
     BFLStatus,
@@ -30,6 +31,7 @@ import requests
 import torch
 import base64
 import time
+from server import PromptServer
 
 
 def convert_mask_to_image(mask: torch.Tensor):
@@ -42,14 +44,19 @@ def convert_mask_to_image(mask: torch.Tensor):
 
 
 def handle_bfl_synchronous_operation(
-    operation: SynchronousOperation, timeout_bfl_calls=360
+    operation: SynchronousOperation,
+    timeout_bfl_calls=360,
+    node_id: Union[str, None] = None,
 ):
     response_api: BFLFluxProGenerateResponse = operation.execute()
     return _poll_until_generated(
-        response_api.polling_url, timeout=timeout_bfl_calls
+        response_api.polling_url, timeout=timeout_bfl_calls, node_id=node_id
     )
 
-def _poll_until_generated(polling_url: str, timeout=360):
+
+def _poll_until_generated(
+    polling_url: str, timeout=360, node_id: Union[str, None] = None
+):
     # used bfl-comfy-nodes to verify code implementation:
     # https://github.com/black-forest-labs/bfl-comfy-nodes/tree/main
     start_time = time.time()
@@ -61,11 +68,21 @@ def _poll_until_generated(polling_url: str, timeout=360):
     request = requests.Request(method=HttpMethod.GET, url=polling_url)
     # NOTE: should True loop be replaced with checking if workflow has been interrupted?
     while True:
+        if node_id:
+            time_elapsed = time.time() - start_time
+            PromptServer.instance.send_progress_text(
+                f"Generating ({time_elapsed:.0f}s)", node_id
+            )
+
         response = requests.Session().send(request.prepare())
         if response.status_code == 200:
             result = response.json()
             if result["status"] == BFLStatus.ready:
                 img_url = result["result"]["sample"]
+                if node_id:
+                    PromptServer.instance.send_progress_text(
+                        f"Result URL: {img_url}", node_id
+                    )
                 img_response = requests.get(img_url)
                 return process_image_response(img_response)
             elif result["status"] in [
@@ -180,6 +197,7 @@ class FluxProUltraImageNode(ComfyNodeABC):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -212,6 +230,7 @@ class FluxProUltraImageNode(ComfyNodeABC):
         seed=0,
         image_prompt=None,
         image_prompt_strength=0.1,
+        unique_id: Union[str, None] = None,
         **kwargs,
     ):
         if image_prompt is None:
@@ -246,7 +265,7 @@ class FluxProUltraImageNode(ComfyNodeABC):
             ),
             auth_kwargs=kwargs,
         )
-        output_image = handle_bfl_synchronous_operation(operation)
+        output_image = handle_bfl_synchronous_operation(operation, node_id=unique_id)
         return (output_image,)
 
 
@@ -320,6 +339,7 @@ class FluxProImageNode(ComfyNodeABC):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -338,6 +358,7 @@ class FluxProImageNode(ComfyNodeABC):
         seed=0,
         image_prompt=None,
         # image_prompt_strength=0.1,
+        unique_id: Union[str, None] = None,
         **kwargs,
     ):
         image_prompt = (
@@ -363,7 +384,7 @@ class FluxProImageNode(ComfyNodeABC):
             ),
             auth_kwargs=kwargs,
         )
-        output_image = handle_bfl_synchronous_operation(operation)
+        output_image = handle_bfl_synchronous_operation(operation, node_id=unique_id)
         return (output_image,)
 
 
@@ -457,11 +478,11 @@ class FluxProExpandNode(ComfyNodeABC):
                     },
                 ),
             },
-            "optional": {
-            },
+            "optional": {},
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -483,6 +504,7 @@ class FluxProExpandNode(ComfyNodeABC):
         steps: int,
         guidance: float,
         seed=0,
+        unique_id: Union[str, None] = None,
         **kwargs,
     ):
         image = convert_image_to_base64(image)
@@ -508,7 +530,7 @@ class FluxProExpandNode(ComfyNodeABC):
             ),
             auth_kwargs=kwargs,
         )
-        output_image = handle_bfl_synchronous_operation(operation)
+        output_image = handle_bfl_synchronous_operation(operation, node_id=unique_id)
         return (output_image,)
 
 
@@ -568,11 +590,11 @@ class FluxProFillNode(ComfyNodeABC):
                     },
                 ),
             },
-            "optional": {
-            },
+            "optional": {},
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -591,13 +613,14 @@ class FluxProFillNode(ComfyNodeABC):
         steps: int,
         guidance: float,
         seed=0,
+        unique_id: Union[str, None] = None,
         **kwargs,
     ):
         # prepare mask
         mask = resize_mask_to_image(mask, image)
         mask = convert_image_to_base64(convert_mask_to_image(mask))
         # make sure image will have alpha channel removed
-        image = convert_image_to_base64(image[:,:,:,:3])
+        image = convert_image_to_base64(image[:, :, :, :3])
 
         operation = SynchronousOperation(
             endpoint=ApiEndpoint(
@@ -617,7 +640,7 @@ class FluxProFillNode(ComfyNodeABC):
             ),
             auth_kwargs=kwargs,
         )
-        output_image = handle_bfl_synchronous_operation(operation)
+        output_image = handle_bfl_synchronous_operation(operation, node_id=unique_id)
         return (output_image,)
 
 
@@ -702,11 +725,11 @@ class FluxProCannyNode(ComfyNodeABC):
                     },
                 ),
             },
-            "optional": {
-            },
+            "optional": {},
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -727,9 +750,10 @@ class FluxProCannyNode(ComfyNodeABC):
         steps: int,
         guidance: float,
         seed=0,
+        unique_id: Union[str, None] = None,
         **kwargs,
     ):
-        control_image = convert_image_to_base64(control_image[:,:,:,:3])
+        control_image = convert_image_to_base64(control_image[:, :, :, :3])
         preprocessed_image = None
 
         # scale canny threshold between 0-500, to match BFL's API
@@ -765,7 +789,7 @@ class FluxProCannyNode(ComfyNodeABC):
             ),
             auth_kwargs=kwargs,
         )
-        output_image = handle_bfl_synchronous_operation(operation)
+        output_image = handle_bfl_synchronous_operation(operation, node_id=unique_id)
         return (output_image,)
 
 
@@ -830,11 +854,11 @@ class FluxProDepthNode(ComfyNodeABC):
                     },
                 ),
             },
-            "optional": {
-            },
+            "optional": {},
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -853,6 +877,7 @@ class FluxProDepthNode(ComfyNodeABC):
         steps: int,
         guidance: float,
         seed=0,
+        unique_id: Union[str, None] = None,
         **kwargs,
     ):
         control_image = convert_image_to_base64(control_image[:,:,:,:3])
@@ -880,7 +905,7 @@ class FluxProDepthNode(ComfyNodeABC):
             ),
             auth_kwargs=kwargs,
         )
-        output_image = handle_bfl_synchronous_operation(operation)
+        output_image = handle_bfl_synchronous_operation(operation, node_id=unique_id)
         return (output_image,)
 
 

comfy_api_nodes/nodes_ideogram.py

@@ -23,6 +23,7 @@ from comfy_api_nodes.apinode_utils import (
     bytesio_to_image_tensor,
     resize_mask_to_image,
 )
+from server import PromptServer
 
 V1_V1_RES_MAP = {
   "Auto":"AUTO",
@@ -232,6 +233,19 @@ def download_and_process_images(image_urls):
     return stacked_tensors
 
 
+def display_image_urls_on_node(image_urls, node_id):
+    if node_id and image_urls:
+        if len(image_urls) == 1:
+            PromptServer.instance.send_progress_text(
+                f"Generated Image URL:\n{image_urls[0]}", node_id
+            )
+        else:
+            urls_text = "Generated Image URLs:\n" + "\n".join(
+                f"{i+1}. {url}" for i, url in enumerate(image_urls)
+            )
+            PromptServer.instance.send_progress_text(urls_text, node_id)
+
+
 class IdeogramV1(ComfyNodeABC):
     """
     Generates images using the Ideogram V1 model.
@@ -304,6 +318,7 @@ class IdeogramV1(ComfyNodeABC):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -322,6 +337,7 @@ class IdeogramV1(ComfyNodeABC):
         seed=0,
         negative_prompt="",
         num_images=1,
+        unique_id=None,
         **kwargs,
     ):
         # Determine the model based on turbo setting
@@ -361,6 +377,7 @@ class IdeogramV1(ComfyNodeABC):
         if not image_urls:
             raise Exception("No image URLs were generated in the response")
 
+        display_image_urls_on_node(image_urls, unique_id)
         return (download_and_process_images(image_urls),)
 
 
@@ -460,6 +477,7 @@ class IdeogramV2(ComfyNodeABC):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -481,6 +499,7 @@ class IdeogramV2(ComfyNodeABC):
         negative_prompt="",
         num_images=1,
         color_palette="",
+        unique_id=None,
         **kwargs,
     ):
         aspect_ratio = V1_V2_RATIO_MAP.get(aspect_ratio, None)
@@ -534,6 +553,7 @@ class IdeogramV2(ComfyNodeABC):
         if not image_urls:
             raise Exception("No image URLs were generated in the response")
 
+        display_image_urls_on_node(image_urls, unique_id)
         return (download_and_process_images(image_urls),)
 
 class IdeogramV3(ComfyNodeABC):
@@ -623,6 +643,7 @@ class IdeogramV3(ComfyNodeABC):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -643,6 +664,7 @@ class IdeogramV3(ComfyNodeABC):
         seed=0,
         num_images=1,
         rendering_speed="BALANCED",
+        unique_id=None,
         **kwargs,
     ):
         # Check if both image and mask are provided for editing mode
@@ -762,6 +784,7 @@ class IdeogramV3(ComfyNodeABC):
         if not image_urls:
             raise Exception("No image URLs were generated in the response")
 
+        display_image_urls_on_node(image_urls, unique_id)
         return (download_and_process_images(image_urls),)
 
 
@@ -776,4 +799,3 @@ NODE_DISPLAY_NAME_MAPPINGS = {
     "IdeogramV2": "Ideogram V2",
     "IdeogramV3": "Ideogram V3",
 }
-

comfy_api_nodes/nodes_kling.py

@@ -6,6 +6,7 @@ For source of truth on the allowed permutations of request fields, please refere
 
 from __future__ import annotations
 from typing import Optional, TypeVar, Any
+from collections.abc import Callable
 import math
 import logging
 
@@ -64,6 +65,12 @@ from comfy_api_nodes.apinode_utils import (
     download_url_to_image_tensor,
 )
 from comfy_api_nodes.mapper_utils import model_field_to_node_input
+from comfy_api_nodes.util.validation_utils import (
+    validate_image_dimensions,
+    validate_image_aspect_ratio,
+    validate_video_dimensions,
+    validate_video_duration,
+)
 from comfy_api.input.basic_types import AudioInput
 from comfy_api.input.video_types import VideoInput
 from comfy_api.input_impl import VideoFromFile
@@ -79,13 +86,20 @@ PATH_CHARACTER_IMAGE = f"/proxy/kling/{KLING_API_VERSION}/images/generations"
 PATH_VIRTUAL_TRY_ON = f"/proxy/kling/{KLING_API_VERSION}/images/kolors-virtual-try-on"
 PATH_IMAGE_GENERATIONS = f"/proxy/kling/{KLING_API_VERSION}/images/generations"
 
-
 MAX_PROMPT_LENGTH_T2V = 2500
 MAX_PROMPT_LENGTH_I2V = 500
 MAX_PROMPT_LENGTH_IMAGE_GEN = 500
 MAX_NEGATIVE_PROMPT_LENGTH_IMAGE_GEN = 200
 MAX_PROMPT_LENGTH_LIP_SYNC = 120
 
+AVERAGE_DURATION_T2V = 319
+AVERAGE_DURATION_I2V = 164
+AVERAGE_DURATION_LIP_SYNC = 455
+AVERAGE_DURATION_VIRTUAL_TRY_ON = 19
+AVERAGE_DURATION_IMAGE_GEN = 32
+AVERAGE_DURATION_VIDEO_EFFECTS = 320
+AVERAGE_DURATION_VIDEO_EXTEND = 320
+
 R = TypeVar("R")
 
 
@@ -95,7 +109,13 @@ class KlingApiError(Exception):
     pass
 
 
-def poll_until_finished(auth_kwargs: dict[str,str], api_endpoint: ApiEndpoint[Any, R]) -> R:
+def poll_until_finished(
+    auth_kwargs: dict[str, str],
+    api_endpoint: ApiEndpoint[Any, R],
+    result_url_extractor: Optional[Callable[[R], str]] = None,
+    estimated_duration: Optional[int] = None,
+    node_id: Optional[str] = None,
+) -> R:
     """Polls the Kling API endpoint until the task reaches a terminal state, then returns the response."""
     return PollingOperation(
         poll_endpoint=api_endpoint,
@@ -109,6 +129,9 @@ def poll_until_finished(auth_kwargs: dict[str,str], api_endpoint: ApiEndpoint[An
             else None
         ),
         auth_kwargs=auth_kwargs,
+        result_url_extractor=result_url_extractor,
+        estimated_duration=estimated_duration,
+        node_id=node_id,
     ).execute()
 
 
@@ -192,23 +215,8 @@ def validate_input_image(image: torch.Tensor) -> None:
 
     See: https://app.klingai.com/global/dev/document-api/apiReference/model/imageToVideo
     """
-    if len(image.shape) == 4:
-        height, width = image.shape[1], image.shape[2]
-    elif len(image.shape) == 3:
-        height, width = image.shape[0], image.shape[1]
-    else:
-        raise ValueError("Invalid image tensor shape.")
-
-    # Ensure minimum resolution is met
-    if height < 300:
-        raise ValueError("Image height must be at least 300px")
-    if width < 300:
-        raise ValueError("Image width must be at least 300px")
-
-    # Ensure aspect ratio is within acceptable range
-    aspect_ratio = width / height
-    if aspect_ratio < 1 / 2.5 or aspect_ratio > 2.5:
-        raise ValueError("Image aspect ratio must be between 1:2.5 and 2.5:1")
+    validate_image_dimensions(image, min_width=300, min_height=300)
+    validate_image_aspect_ratio(image, min_aspect_ratio=1 / 2.5, max_aspect_ratio=2.5)
 
 
 def get_camera_control_input_config(
@@ -227,7 +235,9 @@ def get_camera_control_input_config(
 
 
 def get_video_from_response(response) -> KlingVideoResult:
-    """Returns the first video object from the Kling video generation task result."""
+    """Returns the first video object from the Kling video generation task result.
+    Will raise an error if the response is not valid.
+    """
     video = response.data.task_result.videos[0]
     logging.info(
         "Kling task %s succeeded. Video URL: %s", response.data.task_id, video.url
@@ -235,12 +245,37 @@ def get_video_from_response(response) -> KlingVideoResult:
     return video
 
 
+def get_video_url_from_response(response) -> Optional[str]:
+    """Returns the first video url from the Kling video generation task result.
+    Will not raise an error if the response is not valid.
+    """
+    if response and is_valid_video_response(response):
+        return str(get_video_from_response(response).url)
+    else:
+        return None
+
+
 def get_images_from_response(response) -> list[KlingImageResult]:
+    """Returns the list of image objects from the Kling image generation task result.
+    Will raise an error if the response is not valid.
+    """
     images = response.data.task_result.images
     logging.info("Kling task %s succeeded. Images: %s", response.data.task_id, images)
     return images
 
 
+def get_images_urls_from_response(response) -> Optional[str]:
+    """Returns the list of image urls from the Kling image generation task result.
+    Will not raise an error if the response is not valid. If there is only one image, returns the url as a string. If there are multiple images, returns a list of urls.
+    """
+    if response and is_valid_image_response(response):
+        images = get_images_from_response(response)
+        image_urls = [str(image.url) for image in images]
+        return "\n".join(image_urls)
+    else:
+        return None
+
+
 def video_result_to_node_output(
     video: KlingVideoResult,
 ) -> tuple[VideoFromFile, str, str]:
@@ -312,6 +347,7 @@ class KlingCameraControls(KlingNodeBase):
     RETURN_TYPES = ("CAMERA_CONTROL",)
     RETURN_NAMES = ("camera_control",)
     FUNCTION = "main"
+    API_NODE = False  # This is just a helper node, it doesn't make an API call
 
     @classmethod
     def VALIDATE_INPUTS(
@@ -421,6 +457,7 @@ class KlingTextToVideoNode(KlingNodeBase):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -428,7 +465,9 @@ class KlingTextToVideoNode(KlingNodeBase):
     RETURN_NAMES = ("VIDEO", "video_id", "duration")
     DESCRIPTION = "Kling Text to Video Node"
 
-    def get_response(self, task_id: str, auth_kwargs: dict[str,str]) -> KlingText2VideoResponse:
+    def get_response(
+        self, task_id: str, auth_kwargs: dict[str, str], node_id: Optional[str] = None
+    ) -> KlingText2VideoResponse:
         return poll_until_finished(
             auth_kwargs,
             ApiEndpoint(
@@ -437,6 +476,9 @@ class KlingTextToVideoNode(KlingNodeBase):
                 request_model=EmptyRequest,
                 response_model=KlingText2VideoResponse,
             ),
+            result_url_extractor=get_video_url_from_response,
+            estimated_duration=AVERAGE_DURATION_T2V,
+            node_id=node_id,
         )
 
     def api_call(
@@ -449,6 +491,7 @@ class KlingTextToVideoNode(KlingNodeBase):
         camera_control: Optional[KlingCameraControl] = None,
         model_name: Optional[str] = None,
         duration: Optional[str] = None,
+        unique_id: Optional[str] = None,
         **kwargs,
     ) -> tuple[VideoFromFile, str, str]:
         validate_prompts(prompt, negative_prompt, MAX_PROMPT_LENGTH_T2V)
@@ -478,7 +521,9 @@ class KlingTextToVideoNode(KlingNodeBase):
         validate_task_creation_response(task_creation_response)
 
         task_id = task_creation_response.data.task_id
-        final_response = self.get_response(task_id, auth_kwargs=kwargs)
+        final_response = self.get_response(
+            task_id, auth_kwargs=kwargs, node_id=unique_id
+        )
         validate_video_result_response(final_response)
 
         video = get_video_from_response(final_response)
@@ -528,6 +573,7 @@ class KlingCameraControlT2VNode(KlingTextToVideoNode):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -540,6 +586,7 @@ class KlingCameraControlT2VNode(KlingTextToVideoNode):
         cfg_scale: float,
         aspect_ratio: str,
         camera_control: Optional[KlingCameraControl] = None,
+        unique_id: Optional[str] = None,
         **kwargs,
     ):
         return super().api_call(
@@ -613,6 +660,7 @@ class KlingImage2VideoNode(KlingNodeBase):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -620,7 +668,9 @@ class KlingImage2VideoNode(KlingNodeBase):
     RETURN_NAMES = ("VIDEO", "video_id", "duration")
     DESCRIPTION = "Kling Image to Video Node"
 
-    def get_response(self, task_id: str, auth_kwargs: dict[str,str]) -> KlingImage2VideoResponse:
+    def get_response(
+        self, task_id: str, auth_kwargs: dict[str, str], node_id: Optional[str] = None
+    ) -> KlingImage2VideoResponse:
         return poll_until_finished(
             auth_kwargs,
             ApiEndpoint(
@@ -629,6 +679,9 @@ class KlingImage2VideoNode(KlingNodeBase):
                 request_model=KlingImage2VideoRequest,
                 response_model=KlingImage2VideoResponse,
             ),
+            result_url_extractor=get_video_url_from_response,
+            estimated_duration=AVERAGE_DURATION_I2V,
+            node_id=node_id,
         )
 
     def api_call(
@@ -643,6 +696,7 @@ class KlingImage2VideoNode(KlingNodeBase):
         duration: str,
         camera_control: Optional[KlingCameraControl] = None,
         end_frame: Optional[torch.Tensor] = None,
+        unique_id: Optional[str] = None,
         **kwargs,
     ) -> tuple[VideoFromFile]:
         validate_prompts(prompt, negative_prompt, MAX_PROMPT_LENGTH_I2V)
@@ -681,7 +735,9 @@ class KlingImage2VideoNode(KlingNodeBase):
         validate_task_creation_response(task_creation_response)
         task_id = task_creation_response.data.task_id
 
-        final_response = self.get_response(task_id, auth_kwargs=kwargs)
+        final_response = self.get_response(
+            task_id, auth_kwargs=kwargs, node_id=unique_id
+        )
         validate_video_result_response(final_response)
 
         video = get_video_from_response(final_response)
@@ -734,6 +790,7 @@ class KlingCameraControlI2VNode(KlingImage2VideoNode):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -747,6 +804,7 @@ class KlingCameraControlI2VNode(KlingImage2VideoNode):
         cfg_scale: float,
         aspect_ratio: str,
         camera_control: KlingCameraControl,
+        unique_id: Optional[str] = None,
         **kwargs,
     ):
         return super().api_call(
@@ -759,6 +817,7 @@ class KlingCameraControlI2VNode(KlingImage2VideoNode):
             prompt=prompt,
             negative_prompt=negative_prompt,
             camera_control=camera_control,
+            unique_id=unique_id,
             **kwargs,
         )
 
@@ -830,6 +889,7 @@ class KlingStartEndFrameNode(KlingImage2VideoNode):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -844,6 +904,7 @@ class KlingStartEndFrameNode(KlingImage2VideoNode):
         cfg_scale: float,
         aspect_ratio: str,
         mode: str,
+        unique_id: Optional[str] = None,
         **kwargs,
     ):
         mode, duration, model_name = KlingStartEndFrameNode.get_mode_string_mapping()[
@@ -859,6 +920,7 @@ class KlingStartEndFrameNode(KlingImage2VideoNode):
             aspect_ratio=aspect_ratio,
             duration=duration,
             end_frame=end_frame,
+            unique_id=unique_id,
             **kwargs,
         )
 
@@ -892,6 +954,7 @@ class KlingVideoExtendNode(KlingNodeBase):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -899,7 +962,9 @@ class KlingVideoExtendNode(KlingNodeBase):
     RETURN_NAMES = ("VIDEO", "video_id", "duration")
     DESCRIPTION = "Kling Video Extend Node. Extend videos made by other Kling nodes. The video_id is created by using other Kling Nodes."
 
-    def get_response(self, task_id: str, auth_kwargs: dict[str,str]) -> KlingVideoExtendResponse:
+    def get_response(
+        self, task_id: str, auth_kwargs: dict[str, str], node_id: Optional[str] = None
+    ) -> KlingVideoExtendResponse:
         return poll_until_finished(
             auth_kwargs,
             ApiEndpoint(
@@ -908,6 +973,9 @@ class KlingVideoExtendNode(KlingNodeBase):
                 request_model=EmptyRequest,
                 response_model=KlingVideoExtendResponse,
             ),
+            result_url_extractor=get_video_url_from_response,
+            estimated_duration=AVERAGE_DURATION_VIDEO_EXTEND,
+            node_id=node_id,
         )
 
     def api_call(
@@ -916,6 +984,7 @@ class KlingVideoExtendNode(KlingNodeBase):
         negative_prompt: str,
         cfg_scale: float,
         video_id: str,
+        unique_id: Optional[str] = None,
         **kwargs,
     ) -> tuple[VideoFromFile, str, str]:
         validate_prompts(prompt, negative_prompt, MAX_PROMPT_LENGTH_T2V)
@@ -939,7 +1008,9 @@ class KlingVideoExtendNode(KlingNodeBase):
         validate_task_creation_response(task_creation_response)
         task_id = task_creation_response.data.task_id
 
-        final_response = self.get_response(task_id, auth_kwargs=kwargs)
+        final_response = self.get_response(
+            task_id, auth_kwargs=kwargs, node_id=unique_id
+        )
         validate_video_result_response(final_response)
 
         video = get_video_from_response(final_response)
@@ -952,7 +1023,9 @@ class KlingVideoEffectsBase(KlingNodeBase):
     RETURN_TYPES = ("VIDEO", "STRING", "STRING")
     RETURN_NAMES = ("VIDEO", "video_id", "duration")
 
-    def get_response(self, task_id: str, auth_kwargs: dict[str,str]) -> KlingVideoEffectsResponse:
+    def get_response(
+        self, task_id: str, auth_kwargs: dict[str, str], node_id: Optional[str] = None
+    ) -> KlingVideoEffectsResponse:
         return poll_until_finished(
             auth_kwargs,
             ApiEndpoint(
@@ -961,6 +1034,9 @@ class KlingVideoEffectsBase(KlingNodeBase):
                 request_model=EmptyRequest,
                 response_model=KlingVideoEffectsResponse,
             ),
+            result_url_extractor=get_video_url_from_response,
+            estimated_duration=AVERAGE_DURATION_VIDEO_EFFECTS,
+            node_id=node_id,
         )
 
     def api_call(
@@ -972,6 +1048,7 @@ class KlingVideoEffectsBase(KlingNodeBase):
         image_1: torch.Tensor,
         image_2: Optional[torch.Tensor] = None,
         mode: Optional[KlingVideoGenMode] = None,
+        unique_id: Optional[str] = None,
         **kwargs,
     ):
         if dual_character:
@@ -1009,7 +1086,9 @@ class KlingVideoEffectsBase(KlingNodeBase):
         validate_task_creation_response(task_creation_response)
         task_id = task_creation_response.data.task_id
 
-        final_response = self.get_response(task_id, auth_kwargs=kwargs)
+        final_response = self.get_response(
+            task_id, auth_kwargs=kwargs, node_id=unique_id
+        )
         validate_video_result_response(final_response)
 
         video = get_video_from_response(final_response)
@@ -1053,6 +1132,7 @@ class KlingDualCharacterVideoEffectNode(KlingVideoEffectsBase):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -1068,6 +1148,7 @@ class KlingDualCharacterVideoEffectNode(KlingVideoEffectsBase):
         model_name: KlingCharacterEffectModelName,
         mode: KlingVideoGenMode,
         duration: KlingVideoGenDuration,
+        unique_id: Optional[str] = None,
         **kwargs,
     ):
         video, _, duration = super().api_call(
@@ -1078,10 +1159,12 @@ class KlingDualCharacterVideoEffectNode(KlingVideoEffectsBase):
             duration=duration,
             image_1=image_left,
             image_2=image_right,
+            unique_id=unique_id,
             **kwargs,
         )
         return video, duration
 
+
 class KlingSingleImageVideoEffectNode(KlingVideoEffectsBase):
     """Kling Single Image Video Effect Node"""
 
@@ -1117,6 +1200,7 @@ class KlingSingleImageVideoEffectNode(KlingVideoEffectsBase):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -1128,6 +1212,7 @@ class KlingSingleImageVideoEffectNode(KlingVideoEffectsBase):
         effect_scene: KlingSingleImageEffectsScene,
         model_name: KlingSingleImageEffectModelName,
         duration: KlingVideoGenDuration,
+        unique_id: Optional[str] = None,
         **kwargs,
     ):
         return super().api_call(
@@ -1136,6 +1221,7 @@ class KlingSingleImageVideoEffectNode(KlingVideoEffectsBase):
             model_name=model_name,
             duration=duration,
             image_1=image,
+            unique_id=unique_id,
             **kwargs,
         )
 
@@ -1146,6 +1232,17 @@ class KlingLipSyncBase(KlingNodeBase):
     RETURN_TYPES = ("VIDEO", "STRING", "STRING")
     RETURN_NAMES = ("VIDEO", "video_id", "duration")
 
+    def validate_lip_sync_video(self, video: VideoInput):
+        """
+        Validates the input video adheres to the expectations of the Kling Lip Sync API:
+        - Video length does not exceed 10s and is not shorter than 2s
+        - Length and width dimensions should both be between 720px and 1920px
+
+        See: https://app.klingai.com/global/dev/document-api/apiReference/model/videoTolip
+        """
+        validate_video_dimensions(video, 720, 1920)
+        validate_video_duration(video, 2, 10)
+
     def validate_text(self, text: str):
         if not text:
             raise ValueError("Text is required")
@@ -1154,7 +1251,9 @@ class KlingLipSyncBase(KlingNodeBase):
                 f"Text is too long. Maximum length is {MAX_PROMPT_LENGTH_LIP_SYNC} characters."
             )
 
-    def get_response(self, task_id: str, auth_kwargs: dict[str,str]) -> KlingLipSyncResponse:
+    def get_response(
+        self, task_id: str, auth_kwargs: dict[str, str], node_id: Optional[str] = None
+    ) -> KlingLipSyncResponse:
         """Polls the Kling API endpoint until the task reaches a terminal state."""
         return poll_until_finished(
             auth_kwargs,
@@ -1164,6 +1263,9 @@ class KlingLipSyncBase(KlingNodeBase):
                 request_model=EmptyRequest,
                 response_model=KlingLipSyncResponse,
             ),
+            result_url_extractor=get_video_url_from_response,
+            estimated_duration=AVERAGE_DURATION_LIP_SYNC,
+            node_id=node_id,
         )
 
     def api_call(
@@ -1175,10 +1277,12 @@ class KlingLipSyncBase(KlingNodeBase):
         text: Optional[str] = None,
         voice_speed: Optional[float] = None,
         voice_id: Optional[str] = None,
-        **kwargs
+        unique_id: Optional[str] = None,
+        **kwargs,
     ) -> tuple[VideoFromFile, str, str]:
         if text:
             self.validate_text(text)
+        self.validate_lip_sync_video(video)
 
         # Upload video to Comfy API and get download URL
         video_url = upload_video_to_comfyapi(video, auth_kwargs=kwargs)
@@ -1217,7 +1321,9 @@ class KlingLipSyncBase(KlingNodeBase):
         validate_task_creation_response(task_creation_response)
         task_id = task_creation_response.data.task_id
 
-        final_response = self.get_response(task_id, auth_kwargs=kwargs)
+        final_response = self.get_response(
+            task_id, auth_kwargs=kwargs, node_id=unique_id
+        )
         validate_video_result_response(final_response)
 
         video = get_video_from_response(final_response)
@@ -1243,16 +1349,18 @@ class KlingLipSyncAudioToVideoNode(KlingLipSyncBase):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
-    DESCRIPTION = "Kling Lip Sync Audio to Video Node. Syncs mouth movements in a video file to the audio content of an audio file."
+    DESCRIPTION = "Kling Lip Sync Audio to Video Node. Syncs mouth movements in a video file to the audio content of an audio file. When using, ensure that the audio contains clearly distinguishable vocals and that the video contains a distinct face. The audio file should not be larger than 5MB. The video file should not be larger than 100MB, should have height/width between 720px and 1920px, and should be between 2s and 10s in length."
 
     def api_call(
         self,
         video: VideoInput,
         audio: AudioInput,
         voice_language: str,
+        unique_id: Optional[str] = None,
         **kwargs,
     ):
         return super().api_call(
@@ -1260,6 +1368,7 @@ class KlingLipSyncAudioToVideoNode(KlingLipSyncBase):
             audio=audio,
             voice_language=voice_language,
             mode="audio2video",
+            unique_id=unique_id,
             **kwargs,
         )
 
@@ -1352,10 +1461,11 @@ class KlingLipSyncTextToVideoNode(KlingLipSyncBase):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
-    DESCRIPTION = "Kling Lip Sync Text to Video Node. Syncs mouth movements in a video file to a text prompt."
+    DESCRIPTION = "Kling Lip Sync Text to Video Node. Syncs mouth movements in a video file to a text prompt. The video file should not be larger than 100MB, should have height/width between 720px and 1920px, and should be between 2s and 10s in length."
 
     def api_call(
         self,
@@ -1363,6 +1473,7 @@ class KlingLipSyncTextToVideoNode(KlingLipSyncBase):
         text: str,
         voice: str,
         voice_speed: float,
+        unique_id: Optional[str] = None,
         **kwargs,
     ):
         voice_id, voice_language = KlingLipSyncTextToVideoNode.get_voice_config()[voice]
@@ -1373,6 +1484,7 @@ class KlingLipSyncTextToVideoNode(KlingLipSyncBase):
             voice_id=voice_id,
             voice_speed=voice_speed,
             mode="text2video",
+            unique_id=unique_id,
             **kwargs,
         )
 
@@ -1413,13 +1525,14 @@ class KlingVirtualTryOnNode(KlingImageGenerationBase):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
-    DESCRIPTION = "Kling Virtual Try On Node. Input a human image and a cloth image to try on the cloth on the human."
+    DESCRIPTION = "Kling Virtual Try On Node. Input a human image and a cloth image to try on the cloth on the human. You can merge multiple clothing item pictures into one image with a white background."
 
     def get_response(
-        self, task_id: str, auth_kwargs: dict[str,str] = None
+        self, task_id: str, auth_kwargs: dict[str, str], node_id: Optional[str] = None
     ) -> KlingVirtualTryOnResponse:
         return poll_until_finished(
             auth_kwargs,
@@ -1429,6 +1542,9 @@ class KlingVirtualTryOnNode(KlingImageGenerationBase):
                 request_model=EmptyRequest,
                 response_model=KlingVirtualTryOnResponse,
             ),
+            result_url_extractor=get_images_urls_from_response,
+            estimated_duration=AVERAGE_DURATION_VIRTUAL_TRY_ON,
+            node_id=node_id,
         )
 
     def api_call(
@@ -1436,6 +1552,7 @@ class KlingVirtualTryOnNode(KlingImageGenerationBase):
         human_image: torch.Tensor,
         cloth_image: torch.Tensor,
         model_name: KlingVirtualTryOnModelName,
+        unique_id: Optional[str] = None,
         **kwargs,
     ):
         initial_operation = SynchronousOperation(
@@ -1457,7 +1574,9 @@ class KlingVirtualTryOnNode(KlingImageGenerationBase):
         validate_task_creation_response(task_creation_response)
         task_id = task_creation_response.data.task_id
 
-        final_response = self.get_response(task_id, auth_kwargs=kwargs)
+        final_response = self.get_response(
+            task_id, auth_kwargs=kwargs, node_id=unique_id
+        )
         validate_image_result_response(final_response)
 
         images = get_images_from_response(final_response)
@@ -1528,13 +1647,17 @@ class KlingImageGenerationNode(KlingImageGenerationBase):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
     DESCRIPTION = "Kling Image Generation Node. Generate an image from a text prompt with an optional reference image."
 
     def get_response(
-        self, task_id: str, auth_kwargs: Optional[dict[str,str]] = None
+        self,
+        task_id: str,
+        auth_kwargs: Optional[dict[str, str]],
+        node_id: Optional[str] = None,
     ) -> KlingImageGenerationsResponse:
         return poll_until_finished(
             auth_kwargs,
@@ -1544,6 +1667,9 @@ class KlingImageGenerationNode(KlingImageGenerationBase):
                 request_model=EmptyRequest,
                 response_model=KlingImageGenerationsResponse,
             ),
+            result_url_extractor=get_images_urls_from_response,
+            estimated_duration=AVERAGE_DURATION_IMAGE_GEN,
+            node_id=node_id,
         )
 
     def api_call(
@@ -1557,6 +1683,7 @@ class KlingImageGenerationNode(KlingImageGenerationBase):
         n: int,
         aspect_ratio: KlingImageGenAspectRatio,
         image: Optional[torch.Tensor] = None,
+        unique_id: Optional[str] = None,
         **kwargs,
     ):
         self.validate_prompt(prompt, negative_prompt)
@@ -1589,7 +1716,9 @@ class KlingImageGenerationNode(KlingImageGenerationBase):
         validate_task_creation_response(task_creation_response)
         task_id = task_creation_response.data.task_id
 
-        final_response = self.get_response(task_id, auth_kwargs=kwargs)
+        final_response = self.get_response(
+            task_id, auth_kwargs=kwargs, node_id=unique_id
+        )
         validate_image_result_response(final_response)
 
         images = get_images_from_response(final_response)

comfy_api_nodes/nodes_luma.py

@@ -36,11 +36,20 @@ from comfy_api_nodes.apinode_utils import (
     process_image_response,
     validate_string,
 )
+from server import PromptServer
 
 import requests
 import torch
 from io import BytesIO
 
+LUMA_T2V_AVERAGE_DURATION = 105
+LUMA_I2V_AVERAGE_DURATION = 100
+
+def image_result_url_extractor(response: LumaGeneration):
+    return response.assets.image if hasattr(response, "assets") and hasattr(response.assets, "image") else None
+
+def video_result_url_extractor(response: LumaGeneration):
+    return response.assets.video if hasattr(response, "assets") and hasattr(response.assets, "video") else None
 
 class LumaReferenceNode(ComfyNodeABC):
     """
@@ -204,6 +213,7 @@ class LumaImageGenerationNode(ComfyNodeABC):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -217,6 +227,7 @@ class LumaImageGenerationNode(ComfyNodeABC):
         image_luma_ref: LumaReferenceChain = None,
         style_image: torch.Tensor = None,
         character_image: torch.Tensor = None,
+        unique_id: str = None,
         **kwargs,
     ):
         validate_string(prompt, strip_whitespace=True, min_length=3)
@@ -271,6 +282,8 @@ class LumaImageGenerationNode(ComfyNodeABC):
             completed_statuses=[LumaState.completed],
             failed_statuses=[LumaState.failed],
             status_extractor=lambda x: x.state,
+            result_url_extractor=image_result_url_extractor,
+            node_id=unique_id,
             auth_kwargs=kwargs,
         )
         response_poll = operation.execute()
@@ -353,6 +366,7 @@ class LumaImageModifyNode(ComfyNodeABC):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -363,6 +377,7 @@ class LumaImageModifyNode(ComfyNodeABC):
         image: torch.Tensor,
         image_weight: float,
         seed,
+        unique_id: str = None,
         **kwargs,
     ):
         # first, upload image
@@ -399,6 +414,8 @@ class LumaImageModifyNode(ComfyNodeABC):
             completed_statuses=[LumaState.completed],
             failed_statuses=[LumaState.failed],
             status_extractor=lambda x: x.state,
+            result_url_extractor=image_result_url_extractor,
+            node_id=unique_id,
             auth_kwargs=kwargs,
         )
         response_poll = operation.execute()
@@ -473,6 +490,7 @@ class LumaTextToVideoGenerationNode(ComfyNodeABC):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -486,6 +504,7 @@ class LumaTextToVideoGenerationNode(ComfyNodeABC):
         loop: bool,
         seed,
         luma_concepts: LumaConceptChain = None,
+        unique_id: str = None,
         **kwargs,
     ):
         validate_string(prompt, strip_whitespace=False, min_length=3)
@@ -512,6 +531,9 @@ class LumaTextToVideoGenerationNode(ComfyNodeABC):
         )
         response_api: LumaGeneration = operation.execute()
 
+        if unique_id:
+            PromptServer.instance.send_progress_text(f"Luma video generation started: {response_api.id}", unique_id)
+
         operation = PollingOperation(
             poll_endpoint=ApiEndpoint(
                 path=f"/proxy/luma/generations/{response_api.id}",
@@ -522,6 +544,9 @@ class LumaTextToVideoGenerationNode(ComfyNodeABC):
             completed_statuses=[LumaState.completed],
             failed_statuses=[LumaState.failed],
             status_extractor=lambda x: x.state,
+            result_url_extractor=video_result_url_extractor,
+            node_id=unique_id,
+            estimated_duration=LUMA_T2V_AVERAGE_DURATION,
             auth_kwargs=kwargs,
         )
         response_poll = operation.execute()
@@ -597,6 +622,7 @@ class LumaImageToVideoGenerationNode(ComfyNodeABC):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -611,6 +637,7 @@ class LumaImageToVideoGenerationNode(ComfyNodeABC):
         first_image: torch.Tensor = None,
         last_image: torch.Tensor = None,
         luma_concepts: LumaConceptChain = None,
+        unique_id: str = None,
         **kwargs,
     ):
         if first_image is None and last_image is None:
@@ -642,6 +669,9 @@ class LumaImageToVideoGenerationNode(ComfyNodeABC):
         )
         response_api: LumaGeneration = operation.execute()
 
+        if unique_id:
+            PromptServer.instance.send_progress_text(f"Luma video generation started: {response_api.id}", unique_id)
+
         operation = PollingOperation(
             poll_endpoint=ApiEndpoint(
                 path=f"/proxy/luma/generations/{response_api.id}",
@@ -652,6 +682,9 @@ class LumaImageToVideoGenerationNode(ComfyNodeABC):
             completed_statuses=[LumaState.completed],
             failed_statuses=[LumaState.failed],
             status_extractor=lambda x: x.state,
+            result_url_extractor=video_result_url_extractor,
+            node_id=unique_id,
+            estimated_duration=LUMA_I2V_AVERAGE_DURATION,
             auth_kwargs=kwargs,
         )
         response_poll = operation.execute()

comfy_api_nodes/nodes_minimax.py

@@ -1,3 +1,7 @@
+from typing import Union
+import logging
+import torch
+
 from comfy.comfy_types.node_typing import IO
 from comfy_api.input_impl.video_types import VideoFromFile
 from comfy_api_nodes.apis import (
@@ -20,16 +24,19 @@ from comfy_api_nodes.apinode_utils import (
     upload_images_to_comfyapi,
     validate_string,
 )
+from server import PromptServer
 
-import torch
-import logging
 
+I2V_AVERAGE_DURATION = 114
+T2V_AVERAGE_DURATION = 234
 
 class MinimaxTextToVideoNode:
     """
     Generates videos synchronously based on a prompt, and optional parameters using MiniMax's API.
     """
 
+    AVERAGE_DURATION = T2V_AVERAGE_DURATION
+
     @classmethod
     def INPUT_TYPES(s):
         return {
@@ -68,6 +75,7 @@ class MinimaxTextToVideoNode:
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -85,6 +93,7 @@ class MinimaxTextToVideoNode:
         model="T2V-01",
         image: torch.Tensor=None, # used for ImageToVideo
         subject: torch.Tensor=None, # used for SubjectToVideo
+        unique_id: Union[str, None]=None,
         **kwargs,
     ):
         '''
@@ -138,6 +147,8 @@ class MinimaxTextToVideoNode:
             completed_statuses=["Success"],
             failed_statuses=["Fail"],
             status_extractor=lambda x: x.status.value,
+            estimated_duration=self.AVERAGE_DURATION,
+            node_id=unique_id,
             auth_kwargs=kwargs,
         )
         task_result = video_generate_operation.execute()
@@ -164,6 +175,12 @@ class MinimaxTextToVideoNode:
                 f"No video was found in the response. Full response: {file_result.model_dump()}"
             )
         logging.info(f"Generated video URL: {file_url}")
+        if unique_id:
+            if hasattr(file_result.file, "backup_download_url"):
+                message = f"Result URL: {file_url}\nBackup URL: {file_result.file.backup_download_url}"
+            else:
+                message = f"Result URL: {file_url}"
+            PromptServer.instance.send_progress_text(message, unique_id)
 
         video_io = download_url_to_bytesio(file_url)
         if video_io is None:
@@ -178,6 +195,8 @@ class MinimaxImageToVideoNode(MinimaxTextToVideoNode):
     Generates videos synchronously based on an image and prompt, and optional parameters using MiniMax's API.
     """
 
+    AVERAGE_DURATION = I2V_AVERAGE_DURATION
+
     @classmethod
     def INPUT_TYPES(s):
         return {
@@ -223,6 +242,7 @@ class MinimaxImageToVideoNode(MinimaxTextToVideoNode):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -239,6 +259,8 @@ class MinimaxSubjectToVideoNode(MinimaxTextToVideoNode):
     Generates videos synchronously based on an image and prompt, and optional parameters using MiniMax's API.
     """
 
+    AVERAGE_DURATION = T2V_AVERAGE_DURATION
+
     @classmethod
     def INPUT_TYPES(s):
         return {
@@ -282,6 +304,7 @@ class MinimaxSubjectToVideoNode(MinimaxTextToVideoNode):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 

comfy_api_nodes/nodes_openai.py

@@ -96,6 +96,7 @@ class OpenAIDalle2(ComfyNodeABC):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -113,6 +114,7 @@ class OpenAIDalle2(ComfyNodeABC):
         mask=None,
         n=1,
         size="1024x1024",
+        unique_id=None,
         **kwargs
     ):
         validate_string(prompt, strip_whitespace=False)
@@ -176,7 +178,7 @@ class OpenAIDalle2(ComfyNodeABC):
 
         response = operation.execute()
 
-        img_tensor = validate_and_cast_response(response)
+        img_tensor = validate_and_cast_response(response, node_id=unique_id)
         return (img_tensor,)
 
 
@@ -242,6 +244,7 @@ class OpenAIDalle3(ComfyNodeABC):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -258,6 +261,7 @@ class OpenAIDalle3(ComfyNodeABC):
         style="natural",
         quality="standard",
         size="1024x1024",
+        unique_id=None,
         **kwargs
     ):
         validate_string(prompt, strip_whitespace=False)
@@ -284,7 +288,7 @@ class OpenAIDalle3(ComfyNodeABC):
 
         response = operation.execute()
 
-        img_tensor = validate_and_cast_response(response)
+        img_tensor = validate_and_cast_response(response, node_id=unique_id)
         return (img_tensor,)
 
 
@@ -375,6 +379,7 @@ class OpenAIGPTImage1(ComfyNodeABC):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -394,6 +399,7 @@ class OpenAIGPTImage1(ComfyNodeABC):
         mask=None,
         n=1,
         size="1024x1024",
+        unique_id=None,
         **kwargs
     ):
         validate_string(prompt, strip_whitespace=False)
@@ -476,7 +482,7 @@ class OpenAIGPTImage1(ComfyNodeABC):
 
         response = operation.execute()
 
-        img_tensor = validate_and_cast_response(response)
+        img_tensor = validate_and_cast_response(response, node_id=unique_id)
         return (img_tensor,)
 
 

comfy_api_nodes/nodes_pika.py

@@ -121,7 +121,10 @@ class PikaNodeBase(ComfyNodeABC):
     RETURN_TYPES = ("VIDEO",)
 
     def poll_for_task_status(
-        self, task_id: str, auth_kwargs: Optional[dict[str,str]] = None
+        self,
+        task_id: str,
+        auth_kwargs: Optional[dict[str, str]] = None,
+        node_id: Optional[str] = None,
     ) -> PikaGenerateResponse:
         polling_operation = PollingOperation(
             poll_endpoint=ApiEndpoint(
@@ -141,13 +144,19 @@ class PikaNodeBase(ComfyNodeABC):
                 response.progress if hasattr(response, "progress") else None
             ),
             auth_kwargs=auth_kwargs,
+            result_url_extractor=lambda response: (
+                response.url if hasattr(response, "url") else None
+            ),
+            node_id=node_id,
+            estimated_duration=60
         )
         return polling_operation.execute()
 
     def execute_task(
         self,
         initial_operation: SynchronousOperation[R, PikaGenerateResponse],
-        auth_kwargs: Optional[dict[str,str]] = None,
+        auth_kwargs: Optional[dict[str, str]] = None,
+        node_id: Optional[str] = None,
     ) -> tuple[VideoFromFile]:
         """Executes the initial operation then polls for the task status until it is completed.
 
@@ -208,7 +217,8 @@ class PikaImageToVideoV2_2(PikaNodeBase):
         seed: int,
         resolution: str,
         duration: int,
-        **kwargs
+        unique_id: str,
+        **kwargs,
     ) -> tuple[VideoFromFile]:
         # Convert image to BytesIO
         image_bytes_io = tensor_to_bytesio(image)
@@ -238,7 +248,7 @@ class PikaImageToVideoV2_2(PikaNodeBase):
             auth_kwargs=kwargs,
         )
 
-        return self.execute_task(initial_operation, auth_kwargs=kwargs)
+        return self.execute_task(initial_operation, auth_kwargs=kwargs, node_id=unique_id)
 
 
 class PikaTextToVideoNodeV2_2(PikaNodeBase):
@@ -262,6 +272,7 @@ class PikaTextToVideoNodeV2_2(PikaNodeBase):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -275,6 +286,7 @@ class PikaTextToVideoNodeV2_2(PikaNodeBase):
         resolution: str,
         duration: int,
         aspect_ratio: float,
+        unique_id: str,
         **kwargs,
     ) -> tuple[VideoFromFile]:
         initial_operation = SynchronousOperation(
@@ -296,7 +308,7 @@ class PikaTextToVideoNodeV2_2(PikaNodeBase):
             content_type="application/x-www-form-urlencoded",
         )
 
-        return self.execute_task(initial_operation, auth_kwargs=kwargs)
+        return self.execute_task(initial_operation, auth_kwargs=kwargs, node_id=unique_id)
 
 
 class PikaScenesV2_2(PikaNodeBase):
@@ -340,6 +352,7 @@ class PikaScenesV2_2(PikaNodeBase):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -354,6 +367,7 @@ class PikaScenesV2_2(PikaNodeBase):
         duration: int,
         ingredients_mode: str,
         aspect_ratio: float,
+        unique_id: str,
         image_ingredient_1: Optional[torch.Tensor] = None,
         image_ingredient_2: Optional[torch.Tensor] = None,
         image_ingredient_3: Optional[torch.Tensor] = None,
@@ -403,7 +417,7 @@ class PikaScenesV2_2(PikaNodeBase):
             auth_kwargs=kwargs,
         )
 
-        return self.execute_task(initial_operation, auth_kwargs=kwargs)
+        return self.execute_task(initial_operation, auth_kwargs=kwargs, node_id=unique_id)
 
 
 class PikAdditionsNode(PikaNodeBase):
@@ -439,6 +453,7 @@ class PikAdditionsNode(PikaNodeBase):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -451,6 +466,7 @@ class PikAdditionsNode(PikaNodeBase):
         prompt_text: str,
         negative_prompt: str,
         seed: int,
+        unique_id: str,
         **kwargs,
     ) -> tuple[VideoFromFile]:
         # Convert video to BytesIO
@@ -487,7 +503,7 @@ class PikAdditionsNode(PikaNodeBase):
             auth_kwargs=kwargs,
         )
 
-        return self.execute_task(initial_operation, auth_kwargs=kwargs)
+        return self.execute_task(initial_operation, auth_kwargs=kwargs, node_id=unique_id)
 
 
 class PikaSwapsNode(PikaNodeBase):
@@ -532,6 +548,7 @@ class PikaSwapsNode(PikaNodeBase):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -546,6 +563,7 @@ class PikaSwapsNode(PikaNodeBase):
         prompt_text: str,
         negative_prompt: str,
         seed: int,
+        unique_id: str,
         **kwargs,
     ) -> tuple[VideoFromFile]:
         # Convert video to BytesIO
@@ -592,7 +610,7 @@ class PikaSwapsNode(PikaNodeBase):
             auth_kwargs=kwargs,
         )
 
-        return self.execute_task(initial_operation, auth_kwargs=kwargs)
+        return self.execute_task(initial_operation, auth_kwargs=kwargs, node_id=unique_id)
 
 
 class PikaffectsNode(PikaNodeBase):
@@ -637,6 +655,7 @@ class PikaffectsNode(PikaNodeBase):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -649,6 +668,7 @@ class PikaffectsNode(PikaNodeBase):
         prompt_text: str,
         negative_prompt: str,
         seed: int,
+        unique_id: str,
         **kwargs,
     ) -> tuple[VideoFromFile]:
 
@@ -670,7 +690,7 @@ class PikaffectsNode(PikaNodeBase):
             auth_kwargs=kwargs,
         )
 
-        return self.execute_task(initial_operation, auth_kwargs=kwargs)
+        return self.execute_task(initial_operation, auth_kwargs=kwargs, node_id=unique_id)
 
 
 class PikaStartEndFrameNode2_2(PikaNodeBase):
@@ -689,6 +709,7 @@ class PikaStartEndFrameNode2_2(PikaNodeBase):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -703,6 +724,7 @@ class PikaStartEndFrameNode2_2(PikaNodeBase):
         seed: int,
         resolution: str,
         duration: int,
+        unique_id: str,
         **kwargs,
     ) -> tuple[VideoFromFile]:
 
@@ -733,7 +755,7 @@ class PikaStartEndFrameNode2_2(PikaNodeBase):
             auth_kwargs=kwargs,
         )
 
-        return self.execute_task(initial_operation, auth_kwargs=kwargs)
+        return self.execute_task(initial_operation, auth_kwargs=kwargs, node_id=unique_id)
 
 
 NODE_CLASS_MAPPINGS = {

comfy_api_nodes/nodes_pixverse.py

@@ -1,5 +1,5 @@
 from inspect import cleandoc
-
+from typing import Optional
 from comfy_api_nodes.apis.pixverse_api import (
     PixverseTextVideoRequest,
     PixverseImageVideoRequest,
@@ -34,11 +34,22 @@ import requests
 from io import BytesIO
 
 
+AVERAGE_DURATION_T2V = 32
+AVERAGE_DURATION_I2V = 30
+AVERAGE_DURATION_T2T = 52
+
+
+def get_video_url_from_response(
+    response: PixverseGenerationStatusResponse,
+) -> Optional[str]:
+    if response.Resp is None or response.Resp.url is None:
+        return None
+    return str(response.Resp.url)
+
+
 def upload_image_to_pixverse(image: torch.Tensor, auth_kwargs=None):
     # first, upload image to Pixverse and get image id to use in actual generation call
-    files = {
-        "image": tensor_to_bytesio(image)
-    }
+    files = {"image": tensor_to_bytesio(image)}
     operation = SynchronousOperation(
         endpoint=ApiEndpoint(
             path="/proxy/pixverse/image/upload",
@@ -54,7 +65,9 @@ def upload_image_to_pixverse(image: torch.Tensor, auth_kwargs=None):
     response_upload: PixverseImageUploadResponse = operation.execute()
 
     if response_upload.Resp is None:
-        raise Exception(f"PixVerse image upload request failed: '{response_upload.ErrMsg}'")
+        raise Exception(
+            f"PixVerse image upload request failed: '{response_upload.ErrMsg}'"
+        )
 
     return response_upload.Resp.img_id
 
@@ -73,7 +86,7 @@ class PixverseTemplateNode:
     def INPUT_TYPES(s):
         return {
             "required": {
-                "template": (list(pixverse_templates.keys()), ),
+                "template": (list(pixverse_templates.keys()),),
             }
         }
 
@@ -87,7 +100,7 @@ class PixverseTemplateNode:
 
 class PixverseTextToVideoNode(ComfyNodeABC):
     """
-    Generates videos synchronously based on prompt and output_size.
+    Generates videos based on prompt and output_size.
     """
 
     RETURN_TYPES = (IO.VIDEO,)
@@ -108,9 +121,7 @@ class PixverseTextToVideoNode(ComfyNodeABC):
                         "tooltip": "Prompt for the video generation",
                     },
                 ),
-                "aspect_ratio": (
-                    [ratio.value for ratio in PixverseAspectRatio],
-                ),
+                "aspect_ratio": ([ratio.value for ratio in PixverseAspectRatio],),
                 "quality": (
                     [resolution.value for resolution in PixverseQuality],
                     {
@@ -143,12 +154,13 @@ class PixverseTextToVideoNode(ComfyNodeABC):
                     PixverseIO.TEMPLATE,
                     {
                         "tooltip": "An optional template to influence style of generation, created by the PixVerse Template node."
-                    }
-                )
+                    },
+                ),
             },
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -160,8 +172,9 @@ class PixverseTextToVideoNode(ComfyNodeABC):
         duration_seconds: int,
         motion_mode: str,
         seed,
-        negative_prompt: str=None,
-        pixverse_template: int=None,
+        negative_prompt: str = None,
+        pixverse_template: int = None,
+        unique_id: Optional[str] = None,
         **kwargs,
     ):
         validate_string(prompt, strip_whitespace=False)
@@ -205,19 +218,27 @@ class PixverseTextToVideoNode(ComfyNodeABC):
                 response_model=PixverseGenerationStatusResponse,
             ),
             completed_statuses=[PixverseStatus.successful],
-            failed_statuses=[PixverseStatus.contents_moderation, PixverseStatus.failed, PixverseStatus.deleted],
+            failed_statuses=[
+                PixverseStatus.contents_moderation,
+                PixverseStatus.failed,
+                PixverseStatus.deleted,
+            ],
             status_extractor=lambda x: x.Resp.status,
             auth_kwargs=kwargs,
+            node_id=unique_id,
+            result_url_extractor=get_video_url_from_response,
+            estimated_duration=AVERAGE_DURATION_T2V,
         )
         response_poll = operation.execute()
 
         vid_response = requests.get(response_poll.Resp.url)
+
         return (VideoFromFile(BytesIO(vid_response.content)),)
 
 
 class PixverseImageToVideoNode(ComfyNodeABC):
     """
-    Generates videos synchronously based on prompt and output_size.
+    Generates videos based on prompt and output_size.
     """
 
     RETURN_TYPES = (IO.VIDEO,)
@@ -230,9 +251,7 @@ class PixverseImageToVideoNode(ComfyNodeABC):
     def INPUT_TYPES(s):
         return {
             "required": {
-                "image": (
-                    IO.IMAGE,
-                ),
+                "image": (IO.IMAGE,),
                 "prompt": (
                     IO.STRING,
                     {
@@ -273,12 +292,13 @@ class PixverseImageToVideoNode(ComfyNodeABC):
                     PixverseIO.TEMPLATE,
                     {
                         "tooltip": "An optional template to influence style of generation, created by the PixVerse Template node."
-                    }
-                )
+                    },
+                ),
             },
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -290,8 +310,9 @@ class PixverseImageToVideoNode(ComfyNodeABC):
         duration_seconds: int,
         motion_mode: str,
         seed,
-        negative_prompt: str=None,
-        pixverse_template: int=None,
+        negative_prompt: str = None,
+        pixverse_template: int = None,
+        unique_id: Optional[str] = None,
         **kwargs,
     ):
         validate_string(prompt, strip_whitespace=False)
@@ -337,9 +358,16 @@ class PixverseImageToVideoNode(ComfyNodeABC):
                 response_model=PixverseGenerationStatusResponse,
             ),
             completed_statuses=[PixverseStatus.successful],
-            failed_statuses=[PixverseStatus.contents_moderation, PixverseStatus.failed, PixverseStatus.deleted],
+            failed_statuses=[
+                PixverseStatus.contents_moderation,
+                PixverseStatus.failed,
+                PixverseStatus.deleted,
+            ],
             status_extractor=lambda x: x.Resp.status,
             auth_kwargs=kwargs,
+            node_id=unique_id,
+            result_url_extractor=get_video_url_from_response,
+            estimated_duration=AVERAGE_DURATION_I2V,
         )
         response_poll = operation.execute()
 
@@ -349,7 +377,7 @@ class PixverseImageToVideoNode(ComfyNodeABC):
 
 class PixverseTransitionVideoNode(ComfyNodeABC):
     """
-    Generates videos synchronously based on prompt and output_size.
+    Generates videos based on prompt and output_size.
     """
 
     RETURN_TYPES = (IO.VIDEO,)
@@ -362,12 +390,8 @@ class PixverseTransitionVideoNode(ComfyNodeABC):
     def INPUT_TYPES(s):
         return {
             "required": {
-                "first_frame": (
-                    IO.IMAGE,
-                ),
-                "last_frame": (
-                    IO.IMAGE,
-                ),
+                "first_frame": (IO.IMAGE,),
+                "last_frame": (IO.IMAGE,),
                 "prompt": (
                     IO.STRING,
                     {
@@ -408,6 +432,7 @@ class PixverseTransitionVideoNode(ComfyNodeABC):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -420,7 +445,8 @@ class PixverseTransitionVideoNode(ComfyNodeABC):
         duration_seconds: int,
         motion_mode: str,
         seed,
-        negative_prompt: str=None,
+        negative_prompt: str = None,
+        unique_id: Optional[str] = None,
         **kwargs,
     ):
         validate_string(prompt, strip_whitespace=False)
@@ -467,9 +493,16 @@ class PixverseTransitionVideoNode(ComfyNodeABC):
                 response_model=PixverseGenerationStatusResponse,
             ),
             completed_statuses=[PixverseStatus.successful],
-            failed_statuses=[PixverseStatus.contents_moderation, PixverseStatus.failed, PixverseStatus.deleted],
+            failed_statuses=[
+                PixverseStatus.contents_moderation,
+                PixverseStatus.failed,
+                PixverseStatus.deleted,
+            ],
             status_extractor=lambda x: x.Resp.status,
             auth_kwargs=kwargs,
+            node_id=unique_id,
+            result_url_extractor=get_video_url_from_response,
+            estimated_duration=AVERAGE_DURATION_T2V,
         )
         response_poll = operation.execute()
 

comfy_api_nodes/nodes_recraft.py

@@ -1,5 +1,6 @@
 from __future__ import annotations
 from inspect import cleandoc
+from typing import Optional
 from comfy.utils import ProgressBar
 from comfy_extras.nodes_images import SVG # Added
 from comfy.comfy_types.node_typing import IO
@@ -29,6 +30,8 @@ from comfy_api_nodes.apinode_utils import (
     resize_mask_to_image,
     validate_string,
 )
+from server import PromptServer
+
 import torch
 from io import BytesIO
 from PIL import UnidentifiedImageError
@@ -388,6 +391,7 @@ class RecraftTextToImageNode:
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -400,6 +404,7 @@ class RecraftTextToImageNode:
         recraft_style: RecraftStyle = None,
         negative_prompt: str = None,
         recraft_controls: RecraftControls = None,
+        unique_id: Optional[str] = None,
         **kwargs,
     ):
         validate_string(prompt, strip_whitespace=False, max_length=1000)
@@ -436,8 +441,15 @@ class RecraftTextToImageNode:
         )
         response: RecraftImageGenerationResponse = operation.execute()
         images = []
+        urls = []
         for data in response.data:
             with handle_recraft_image_output():
+                if unique_id and data.url:
+                    urls.append(data.url)
+                    urls_string = '\n'.join(urls)
+                    PromptServer.instance.send_progress_text(
+                        f"Result URL: {urls_string}", unique_id
+                    )
                 image = bytesio_to_image_tensor(
                     download_url_to_bytesio(data.url, timeout=1024)
                 )
@@ -763,6 +775,7 @@ class RecraftTextToVectorNode:
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -775,6 +788,7 @@ class RecraftTextToVectorNode:
         seed,
         negative_prompt: str = None,
         recraft_controls: RecraftControls = None,
+        unique_id: Optional[str] = None,
         **kwargs,
     ):
         validate_string(prompt, strip_whitespace=False, max_length=1000)
@@ -809,7 +823,14 @@ class RecraftTextToVectorNode:
         )
         response: RecraftImageGenerationResponse = operation.execute()
         svg_data = []
+        urls = []
         for data in response.data:
+            if unique_id and data.url:
+                urls.append(data.url)
+                # Print result on each iteration in case of error
+                PromptServer.instance.send_progress_text(
+                    f"Result URL: {' '.join(urls)}", unique_id
+                )
             svg_data.append(download_url_to_bytesio(data.url, timeout=1024))
 
         return (SVG(svg_data),)

comfy_api_nodes/nodes_veo2.py

@@ -3,6 +3,7 @@ import logging
 import base64
 import requests
 import torch
+from typing import Optional
 
 from comfy.comfy_types.node_typing import IO, ComfyNodeABC
 from comfy_api.input_impl.video_types import VideoFromFile
@@ -24,6 +25,8 @@ from comfy_api_nodes.apinode_utils import (
     tensor_to_base64_string
 )
 
+AVERAGE_DURATION_VIDEO_GEN = 32
+
 def convert_image_to_base64(image: torch.Tensor):
     if image is None:
         return None
@@ -31,6 +34,22 @@ def convert_image_to_base64(image: torch.Tensor):
     scaled_image = downscale_image_tensor(image, total_pixels=2048*2048)
     return tensor_to_base64_string(scaled_image)
 
+
+def get_video_url_from_response(poll_response: Veo2GenVidPollResponse) -> Optional[str]:
+    if (
+        poll_response.response
+        and hasattr(poll_response.response, "videos")
+        and poll_response.response.videos
+        and len(poll_response.response.videos) > 0
+    ):
+        video = poll_response.response.videos[0]
+    else:
+        return None
+    if hasattr(video, "gcsUri") and video.gcsUri:
+        return str(video.gcsUri)
+    return None
+
+
 class VeoVideoGenerationNode(ComfyNodeABC):
     """
     Generates videos from text prompts using Google's Veo API.
@@ -115,6 +134,7 @@ class VeoVideoGenerationNode(ComfyNodeABC):
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
                 "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
             },
         }
 
@@ -134,6 +154,7 @@ class VeoVideoGenerationNode(ComfyNodeABC):
         person_generation="ALLOW",
         seed=0,
         image=None,
+        unique_id: Optional[str] = None,
         **kwargs,
     ):
         # Prepare the instances for the request
@@ -215,7 +236,10 @@ class VeoVideoGenerationNode(ComfyNodeABC):
                 operationName=operation_name
             ),
             auth_kwargs=kwargs,
-            poll_interval=5.0
+            poll_interval=5.0,
+            result_url_extractor=get_video_url_from_response,
+            node_id=unique_id,
+            estimated_duration=AVERAGE_DURATION_VIDEO_GEN,
         )
 
         # Execute the polling operation

comfy_api_nodes/util/validation_utils.py

@@ -0,0 +1,100 @@
+import logging
+from typing import Optional
+
+import torch
+from comfy_api.input.video_types import VideoInput
+
+
+def get_image_dimensions(image: torch.Tensor) -> tuple[int, int]:
+    if len(image.shape) == 4:
+        return image.shape[1], image.shape[2]
+    elif len(image.shape) == 3:
+        return image.shape[0], image.shape[1]
+    else:
+        raise ValueError("Invalid image tensor shape.")
+
+
+def validate_image_dimensions(
+    image: torch.Tensor,
+    min_width: Optional[int] = None,
+    max_width: Optional[int] = None,
+    min_height: Optional[int] = None,
+    max_height: Optional[int] = None,
+):
+    height, width = get_image_dimensions(image)
+
+    if min_width is not None and width < min_width:
+        raise ValueError(f"Image width must be at least {min_width}px, got {width}px")
+    if max_width is not None and width > max_width:
+        raise ValueError(f"Image width must be at most {max_width}px, got {width}px")
+    if min_height is not None and height < min_height:
+        raise ValueError(
+            f"Image height must be at least {min_height}px, got {height}px"
+        )
+    if max_height is not None and height > max_height:
+        raise ValueError(f"Image height must be at most {max_height}px, got {height}px")
+
+
+def validate_image_aspect_ratio(
+    image: torch.Tensor,
+    min_aspect_ratio: Optional[float] = None,
+    max_aspect_ratio: Optional[float] = None,
+):
+    width, height = get_image_dimensions(image)
+    aspect_ratio = width / height
+
+    if min_aspect_ratio is not None and aspect_ratio < min_aspect_ratio:
+        raise ValueError(
+            f"Image aspect ratio must be at least {min_aspect_ratio}, got {aspect_ratio}"
+        )
+    if max_aspect_ratio is not None and aspect_ratio > max_aspect_ratio:
+        raise ValueError(
+            f"Image aspect ratio must be at most {max_aspect_ratio}, got {aspect_ratio}"
+        )
+
+
+def validate_video_dimensions(
+    video: VideoInput,
+    min_width: Optional[int] = None,
+    max_width: Optional[int] = None,
+    min_height: Optional[int] = None,
+    max_height: Optional[int] = None,
+):
+    try:
+        width, height = video.get_dimensions()
+    except Exception as e:
+        logging.error("Error getting dimensions of video: %s", e)
+        return
+
+    if min_width is not None and width < min_width:
+        raise ValueError(f"Video width must be at least {min_width}px, got {width}px")
+    if max_width is not None and width > max_width:
+        raise ValueError(f"Video width must be at most {max_width}px, got {width}px")
+    if min_height is not None and height < min_height:
+        raise ValueError(
+            f"Video height must be at least {min_height}px, got {height}px"
+        )
+    if max_height is not None and height > max_height:
+        raise ValueError(f"Video height must be at most {max_height}px, got {height}px")
+
+
+def validate_video_duration(
+    video: VideoInput,
+    min_duration: Optional[float] = None,
+    max_duration: Optional[float] = None,
+):
+    try:
+        duration = video.get_duration()
+    except Exception as e:
+        logging.error("Error getting duration of video: %s", e)
+        return
+
+    epsilon = 0.0001
+    if min_duration is not None and min_duration - epsilon > duration:
+        raise ValueError(
+            f"Video duration must be at least {min_duration}s, got {duration}s"
+        )
+    if max_duration is not None and duration > max_duration + epsilon:
+        raise ValueError(
+            f"Video duration must be at most {max_duration}s, got {duration}s"
+        )

comfy_extras/nodes_apg.py

@@ -0,0 +1,76 @@
+import torch
+
+def project(v0, v1):
+    v1 = torch.nn.functional.normalize(v1, dim=[-1, -2, -3])
+    v0_parallel = (v0 * v1).sum(dim=[-1, -2, -3], keepdim=True) * v1
+    v0_orthogonal = v0 - v0_parallel
+    return v0_parallel, v0_orthogonal
+
+class APG:
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "model": ("MODEL",),
+                "eta": ("FLOAT", {"default": 1.0, "min": -10.0, "max": 10.0, "step": 0.01, "tooltip": "Controls the scale of the parallel guidance vector. Default CFG behavior at a setting of 1."}),
+                "norm_threshold": ("FLOAT", {"default": 5.0, "min": 0.0, "max": 50.0, "step": 0.1, "tooltip": "Normalize guidance vector to this value, normalization disable at a setting of 0."}),
+                "momentum": ("FLOAT", {"default": 0.0, "min": -5.0, "max": 1.0, "step": 0.01, "tooltip":"Controls a running average of guidance during diffusion, disabled at a setting of 0."}),
+            }
+        }
+    RETURN_TYPES = ("MODEL",)
+    FUNCTION = "patch"
+    CATEGORY = "sampling/custom_sampling"
+
+    def patch(self, model, eta, norm_threshold, momentum):
+        running_avg = 0
+        prev_sigma = None
+
+        def pre_cfg_function(args):
+            nonlocal running_avg, prev_sigma
+
+            if len(args["conds_out"]) == 1: return args["conds_out"]
+
+            cond = args["conds_out"][0]
+            uncond = args["conds_out"][1]
+            sigma = args["sigma"][0]
+            cond_scale = args["cond_scale"]
+
+            if prev_sigma is not None and sigma > prev_sigma:
+                running_avg = 0
+            prev_sigma = sigma
+
+            guidance = cond - uncond
+
+            if momentum != 0:
+                if not torch.is_tensor(running_avg):
+                    running_avg = guidance
+                else:
+                    running_avg = momentum * running_avg + guidance
+                guidance = running_avg
+
+            if norm_threshold > 0:
+                guidance_norm = guidance.norm(p=2, dim=[-1, -2, -3], keepdim=True)
+                scale = torch.minimum(
+                    torch.ones_like(guidance_norm),
+                    norm_threshold / guidance_norm
+                )
+                guidance = guidance * scale
+
+            guidance_parallel, guidance_orthogonal = project(guidance, cond)
+            modified_guidance = guidance_orthogonal + eta * guidance_parallel
+
+            modified_cond = (uncond + modified_guidance) + (cond - uncond) / cond_scale
+
+            return [modified_cond, uncond] + args["conds_out"][2:]
+
+        m = model.clone()
+        m.set_model_sampler_pre_cfg_function(pre_cfg_function)
+        return (m,)
+
+NODE_CLASS_MAPPINGS = {
+    "APG": APG,
+}
+
+NODE_DISPLAY_NAME_MAPPINGS = {
+    "APG": "Adaptive Projected Guidance",
+}

comfy_extras/nodes_audio.py

@@ -1,5 +1,6 @@
 from __future__ import annotations
 
+import av
 import torchaudio
 import torch
 import comfy.model_management
@@ -7,7 +8,6 @@ import folder_paths
 import os
 import io
 import json
-import struct
 import random
 import hashlib
 import node_helpers
@@ -90,60 +90,118 @@ class VAEDecodeAudio:
         return ({"waveform": audio, "sample_rate": 44100}, )
 
 
-def create_vorbis_comment_block(comment_dict, last_block):
-    vendor_string = b'ComfyUI'
-    vendor_length = len(vendor_string)
+def save_audio(self, audio, filename_prefix="ComfyUI", format="flac", prompt=None, extra_pnginfo=None, quality="128k"):
 
-    comments = []
-    for key, value in comment_dict.items():
-        comment = f"{key}={value}".encode('utf-8')
-        comments.append(struct.pack('<I', len(comment)) + comment)
+    filename_prefix += self.prefix_append
+    full_output_folder, filename, counter, subfolder, filename_prefix = folder_paths.get_save_image_path(filename_prefix, self.output_dir)
+    results: list[FileLocator] = []
 
-    user_comment_list_length = len(comments)
-    user_comments = b''.join(comments)
+    # Prepare metadata dictionary
+    metadata = {}
+    if not args.disable_metadata:
+        if prompt is not None:
+            metadata["prompt"] = json.dumps(prompt)
+        if extra_pnginfo is not None:
+            for x in extra_pnginfo:
+                metadata[x] = json.dumps(extra_pnginfo[x])
 
-    comment_data = struct.pack('<I', vendor_length) + vendor_string + struct.pack('<I', user_comment_list_length) + user_comments
-    if last_block:
-        id = b'\x84'
-    else:
-        id = b'\x04'
-    comment_block = id + struct.pack('>I', len(comment_data))[1:] + comment_data
+    # Opus supported sample rates
+    OPUS_RATES = [8000, 12000, 16000, 24000, 48000]
 
-    return comment_block
+    for (batch_number, waveform) in enumerate(audio["waveform"].cpu()):
+        filename_with_batch_num = filename.replace("%batch_num%", str(batch_number))
+        file = f"{filename_with_batch_num}_{counter:05}_.{format}"
+        output_path = os.path.join(full_output_folder, file)
 
-def insert_or_replace_vorbis_comment(flac_io, comment_dict):
-    if len(comment_dict) == 0:
-        return flac_io
+        # Use original sample rate initially
+        sample_rate = audio["sample_rate"]
 
-    flac_io.seek(4)
+        # Handle Opus sample rate requirements
+        if format == "opus":
+            if sample_rate > 48000:
+                sample_rate = 48000
+            elif sample_rate not in OPUS_RATES:
+                # Find the next highest supported rate
+                for rate in sorted(OPUS_RATES):
+                    if rate > sample_rate:
+                        sample_rate = rate
+                        break
+                if sample_rate not in OPUS_RATES:  # Fallback if still not supported
+                    sample_rate = 48000
 
-    blocks = []
-    last_block = False
+            # Resample if necessary
+            if sample_rate != audio["sample_rate"]:
+                waveform = torchaudio.functional.resample(waveform, audio["sample_rate"], sample_rate)
 
-    while not last_block:
-        header = flac_io.read(4)
-        last_block = (header[0] & 0x80) != 0
-        block_type = header[0] & 0x7F
-        block_length = struct.unpack('>I', b'\x00' + header[1:])[0]
-        block_data = flac_io.read(block_length)
+        # Create in-memory WAV buffer
+        wav_buffer = io.BytesIO()
+        torchaudio.save(wav_buffer, waveform, sample_rate, format="WAV")
+        wav_buffer.seek(0)  # Rewind for reading
 
-        if block_type == 4 or block_type == 1:
-            pass
-        else:
-            header = bytes([(header[0] & (~0x80))]) + header[1:]
-            blocks.append(header + block_data)
+        # Use PyAV to convert and add metadata
+        input_container = av.open(wav_buffer)
 
-    blocks.append(create_vorbis_comment_block(comment_dict, last_block=True))
+        # Create output with specified format
+        output_buffer = io.BytesIO()
+        output_container = av.open(output_buffer, mode='w', format=format)
 
-    new_flac_io = io.BytesIO()
-    new_flac_io.write(b'fLaC')
-    for block in blocks:
-        new_flac_io.write(block)
+        # Set metadata on the container
+        for key, value in metadata.items():
+            output_container.metadata[key] = value
 
-    new_flac_io.write(flac_io.read())
-    return new_flac_io
+        # Set up the output stream with appropriate properties
+        input_container.streams.audio[0]
+        if format == "opus":
+            out_stream = output_container.add_stream("libopus", rate=sample_rate)
+            if quality == "64k":
+                out_stream.bit_rate = 64000
+            elif quality == "96k":
+                out_stream.bit_rate = 96000
+            elif quality == "128k":
+                out_stream.bit_rate = 128000
+            elif quality == "192k":
+                out_stream.bit_rate = 192000
+            elif quality == "320k":
+                out_stream.bit_rate = 320000
+        elif format == "mp3":
+            out_stream = output_container.add_stream("libmp3lame", rate=sample_rate)
+            if quality == "V0":
+                #TODO i would really love to support V3 and V5 but there doesn't seem to be a way to set the qscale level, the property below is a bool
+                out_stream.codec_context.qscale = 1
+            elif quality == "128k":
+                out_stream.bit_rate = 128000
+            elif quality == "320k":
+                out_stream.bit_rate = 320000
+        else: #format == "flac":
+            out_stream = output_container.add_stream("flac", rate=sample_rate)
 
 
+        # Copy frames from input to output
+        for frame in input_container.decode(audio=0):
+            frame.pts = None  # Let PyAV handle timestamps
+            output_container.mux(out_stream.encode(frame))
+
+        # Flush encoder
+        output_container.mux(out_stream.encode(None))
+
+        # Close containers
+        output_container.close()
+        input_container.close()
+
+        # Write the output to file
+        output_buffer.seek(0)
+        with open(output_path, 'wb') as f:
+            f.write(output_buffer.getbuffer())
+
+        results.append({
+            "filename": file,
+            "subfolder": subfolder,
+            "type": self.type
+        })
+        counter += 1
+
+    return { "ui": { "audio": results } }
+
 class SaveAudio:
     def __init__(self):
         self.output_dir = folder_paths.get_output_directory()
@@ -153,50 +211,70 @@ class SaveAudio:
     @classmethod
     def INPUT_TYPES(s):
         return {"required": { "audio": ("AUDIO", ),
-                              "filename_prefix": ("STRING", {"default": "audio/ComfyUI"})},
+                            "filename_prefix": ("STRING", {"default": "audio/ComfyUI"}),
+                            },
                 "hidden": {"prompt": "PROMPT", "extra_pnginfo": "EXTRA_PNGINFO"},
                 }
 
     RETURN_TYPES = ()
-    FUNCTION = "save_audio"
+    FUNCTION = "save_flac"
 
     OUTPUT_NODE = True
 
     CATEGORY = "audio"
 
-    def save_audio(self, audio, filename_prefix="ComfyUI", prompt=None, extra_pnginfo=None):
-        filename_prefix += self.prefix_append
-        full_output_folder, filename, counter, subfolder, filename_prefix = folder_paths.get_save_image_path(filename_prefix, self.output_dir)
-        results: list[FileLocator] = []
+    def save_flac(self, audio, filename_prefix="ComfyUI", format="flac", prompt=None, extra_pnginfo=None):
+        return save_audio(self, audio, filename_prefix, format, prompt, extra_pnginfo)
 
-        metadata = {}
-        if not args.disable_metadata:
-            if prompt is not None:
-                metadata["prompt"] = json.dumps(prompt)
-            if extra_pnginfo is not None:
-                for x in extra_pnginfo:
-                    metadata[x] = json.dumps(extra_pnginfo[x])
+class SaveAudioMP3:
+    def __init__(self):
+        self.output_dir = folder_paths.get_output_directory()
+        self.type = "output"
+        self.prefix_append = ""
 
-        for (batch_number, waveform) in enumerate(audio["waveform"].cpu()):
-            filename_with_batch_num = filename.replace("%batch_num%", str(batch_number))
-            file = f"{filename_with_batch_num}_{counter:05}_.flac"
+    @classmethod
+    def INPUT_TYPES(s):
+        return {"required": { "audio": ("AUDIO", ),
+                            "filename_prefix": ("STRING", {"default": "audio/ComfyUI"}),
+                            "quality": (["V0", "128k", "320k"], {"default": "V0"}),
+                            },
+                "hidden": {"prompt": "PROMPT", "extra_pnginfo": "EXTRA_PNGINFO"},
+                }
 
-            buff = io.BytesIO()
-            torchaudio.save(buff, waveform, audio["sample_rate"], format="FLAC")
+    RETURN_TYPES = ()
+    FUNCTION = "save_mp3"
 
-            buff = insert_or_replace_vorbis_comment(buff, metadata)
+    OUTPUT_NODE = True
 
-            with open(os.path.join(full_output_folder, file), 'wb') as f:
-                f.write(buff.getbuffer())
+    CATEGORY = "audio"
 
-            results.append({
-                "filename": file,
-                "subfolder": subfolder,
-                "type": self.type
-            })
-            counter += 1
+    def save_mp3(self, audio, filename_prefix="ComfyUI", format="mp3", prompt=None, extra_pnginfo=None, quality="128k"):
+        return save_audio(self, audio, filename_prefix, format, prompt, extra_pnginfo, quality)
 
-        return { "ui": { "audio": results } }
+class SaveAudioOpus:
+    def __init__(self):
+        self.output_dir = folder_paths.get_output_directory()
+        self.type = "output"
+        self.prefix_append = ""
+
+    @classmethod
+    def INPUT_TYPES(s):
+        return {"required": { "audio": ("AUDIO", ),
+                            "filename_prefix": ("STRING", {"default": "audio/ComfyUI"}),
+                            "quality": (["64k", "96k", "128k", "192k", "320k"], {"default": "128k"}),
+                            },
+                "hidden": {"prompt": "PROMPT", "extra_pnginfo": "EXTRA_PNGINFO"},
+                }
+
+    RETURN_TYPES = ()
+    FUNCTION = "save_opus"
+
+    OUTPUT_NODE = True
+
+    CATEGORY = "audio"
+
+    def save_opus(self, audio, filename_prefix="ComfyUI", format="opus", prompt=None, extra_pnginfo=None, quality="V3"):
+        return save_audio(self, audio, filename_prefix, format, prompt, extra_pnginfo, quality)
 
 class PreviewAudio(SaveAudio):
     def __init__(self):
@@ -248,7 +326,20 @@ NODE_CLASS_MAPPINGS = {
     "VAEEncodeAudio": VAEEncodeAudio,
     "VAEDecodeAudio": VAEDecodeAudio,
     "SaveAudio": SaveAudio,
+    "SaveAudioMP3": SaveAudioMP3,
+    "SaveAudioOpus": SaveAudioOpus,
     "LoadAudio": LoadAudio,
     "PreviewAudio": PreviewAudio,
     "ConditioningStableAudio": ConditioningStableAudio,
 }
+
+NODE_DISPLAY_NAME_MAPPINGS = {
+    "EmptyLatentAudio": "Empty Latent Audio",
+    "VAEEncodeAudio": "VAE Encode Audio",
+    "VAEDecodeAudio": "VAE Decode Audio",
+    "PreviewAudio": "Preview Audio",
+    "LoadAudio": "Load Audio",
+    "SaveAudio": "Save Audio (FLAC)",
+    "SaveAudioMP3": "Save Audio (MP3)",
+    "SaveAudioOpus": "Save Audio (Opus)",
+}

comfy_extras/nodes_camera_trajectory.py

@@ -0,0 +1,218 @@
+import nodes
+import torch
+import numpy as np
+from einops import rearrange
+import comfy.model_management
+
+
+
+MAX_RESOLUTION = nodes.MAX_RESOLUTION
+
+CAMERA_DICT = {
+    "base_T_norm": 1.5,
+    "base_angle": np.pi/3,
+    "Static": {     "angle":[0., 0., 0.],   "T":[0., 0., 0.]},
+    "Pan Up": {     "angle":[0., 0., 0.],   "T":[0., -1., 0.]},
+    "Pan Down": {   "angle":[0., 0., 0.],   "T":[0.,1.,0.]},
+    "Pan Left": {   "angle":[0., 0., 0.],   "T":[-1.,0.,0.]},
+    "Pan Right": {  "angle":[0., 0., 0.],   "T": [1.,0.,0.]},
+    "Zoom In": {    "angle":[0., 0., 0.],   "T": [0.,0.,2.]},
+    "Zoom Out": {   "angle":[0., 0., 0.],   "T": [0.,0.,-2.]},
+    "Anti Clockwise (ACW)": {        "angle": [0., 0., -1.],  "T":[0., 0., 0.]},
+    "ClockWise (CW)": {         "angle": [0., 0., 1.], "T":[0., 0., 0.]},
+}
+
+
+def process_pose_params(cam_params, width=672, height=384, original_pose_width=1280, original_pose_height=720, device='cpu'):
+
+    def get_relative_pose(cam_params):
+        """Copied from https://github.com/hehao13/CameraCtrl/blob/main/inference.py
+        """
+        abs_w2cs = [cam_param.w2c_mat for cam_param in cam_params]
+        abs_c2ws = [cam_param.c2w_mat for cam_param in cam_params]
+        cam_to_origin = 0
+        target_cam_c2w = np.array([
+            [1, 0, 0, 0],
+            [0, 1, 0, -cam_to_origin],
+            [0, 0, 1, 0],
+            [0, 0, 0, 1]
+        ])
+        abs2rel = target_cam_c2w @ abs_w2cs[0]
+        ret_poses = [target_cam_c2w, ] + [abs2rel @ abs_c2w for abs_c2w in abs_c2ws[1:]]
+        ret_poses = np.array(ret_poses, dtype=np.float32)
+        return ret_poses
+
+    """Modified from https://github.com/hehao13/CameraCtrl/blob/main/inference.py
+    """
+    cam_params = [Camera(cam_param) for cam_param in cam_params]
+
+    sample_wh_ratio = width / height
+    pose_wh_ratio = original_pose_width / original_pose_height  # Assuming placeholder ratios, change as needed
+
+    if pose_wh_ratio > sample_wh_ratio:
+        resized_ori_w = height * pose_wh_ratio
+        for cam_param in cam_params:
+            cam_param.fx = resized_ori_w * cam_param.fx / width
+    else:
+        resized_ori_h = width / pose_wh_ratio
+        for cam_param in cam_params:
+            cam_param.fy = resized_ori_h * cam_param.fy / height
+
+    intrinsic = np.asarray([[cam_param.fx * width,
+                            cam_param.fy * height,
+                            cam_param.cx * width,
+                            cam_param.cy * height]
+                            for cam_param in cam_params], dtype=np.float32)
+
+    K = torch.as_tensor(intrinsic)[None]  # [1, 1, 4]
+    c2ws = get_relative_pose(cam_params)  # Assuming this function is defined elsewhere
+    c2ws = torch.as_tensor(c2ws)[None]  # [1, n_frame, 4, 4]
+    plucker_embedding = ray_condition(K, c2ws, height, width, device=device)[0].permute(0, 3, 1, 2).contiguous()  # V, 6, H, W
+    plucker_embedding = plucker_embedding[None]
+    plucker_embedding = rearrange(plucker_embedding, "b f c h w -> b f h w c")[0]
+    return plucker_embedding
+
+class Camera(object):
+    """Copied from https://github.com/hehao13/CameraCtrl/blob/main/inference.py
+    """
+    def __init__(self, entry):
+        fx, fy, cx, cy = entry[1:5]
+        self.fx = fx
+        self.fy = fy
+        self.cx = cx
+        self.cy = cy
+        c2w_mat = np.array(entry[7:]).reshape(4, 4)
+        self.c2w_mat = c2w_mat
+        self.w2c_mat = np.linalg.inv(c2w_mat)
+
+def ray_condition(K, c2w, H, W, device):
+    """Copied from https://github.com/hehao13/CameraCtrl/blob/main/inference.py
+    """
+    # c2w: B, V, 4, 4
+    # K: B, V, 4
+
+    B = K.shape[0]
+
+    j, i = torch.meshgrid(
+        torch.linspace(0, H - 1, H, device=device, dtype=c2w.dtype),
+        torch.linspace(0, W - 1, W, device=device, dtype=c2w.dtype),
+        indexing='ij'
+    )
+    i = i.reshape([1, 1, H * W]).expand([B, 1, H * W]) + 0.5  # [B, HxW]
+    j = j.reshape([1, 1, H * W]).expand([B, 1, H * W]) + 0.5  # [B, HxW]
+
+    fx, fy, cx, cy = K.chunk(4, dim=-1)  # B,V, 1
+
+    zs = torch.ones_like(i)  # [B, HxW]
+    xs = (i - cx) / fx * zs
+    ys = (j - cy) / fy * zs
+    zs = zs.expand_as(ys)
+
+    directions = torch.stack((xs, ys, zs), dim=-1)  # B, V, HW, 3
+    directions = directions / directions.norm(dim=-1, keepdim=True)  # B, V, HW, 3
+
+    rays_d = directions @ c2w[..., :3, :3].transpose(-1, -2)  # B, V, 3, HW
+    rays_o = c2w[..., :3, 3]  # B, V, 3
+    rays_o = rays_o[:, :, None].expand_as(rays_d)  # B, V, 3, HW
+    # c2w @ dirctions
+    rays_dxo = torch.cross(rays_o, rays_d)
+    plucker = torch.cat([rays_dxo, rays_d], dim=-1)
+    plucker = plucker.reshape(B, c2w.shape[1], H, W, 6)  # B, V, H, W, 6
+    # plucker = plucker.permute(0, 1, 4, 2, 3)
+    return plucker
+
+def get_camera_motion(angle, T, speed, n=81):
+    def compute_R_form_rad_angle(angles):
+        theta_x, theta_y, theta_z = angles
+        Rx = np.array([[1, 0, 0],
+                    [0, np.cos(theta_x), -np.sin(theta_x)],
+                    [0, np.sin(theta_x), np.cos(theta_x)]])
+
+        Ry = np.array([[np.cos(theta_y), 0, np.sin(theta_y)],
+                    [0, 1, 0],
+                    [-np.sin(theta_y), 0, np.cos(theta_y)]])
+
+        Rz = np.array([[np.cos(theta_z), -np.sin(theta_z), 0],
+                    [np.sin(theta_z), np.cos(theta_z), 0],
+                    [0, 0, 1]])
+
+        R = np.dot(Rz, np.dot(Ry, Rx))
+        return R
+    RT = []
+    for i in range(n):
+        _angle = (i/n)*speed*(CAMERA_DICT["base_angle"])*angle
+        R = compute_R_form_rad_angle(_angle)
+        _T=(i/n)*speed*(CAMERA_DICT["base_T_norm"])*(T.reshape(3,1))
+        _RT = np.concatenate([R,_T], axis=1)
+        RT.append(_RT)
+    RT = np.stack(RT)
+    return RT
+
+class WanCameraEmbedding:
+    @classmethod
+    def INPUT_TYPES(cls):
+        return {
+            "required": {
+                "camera_pose":(["Static","Pan Up","Pan Down","Pan Left","Pan Right","Zoom In","Zoom Out","Anti Clockwise (ACW)", "ClockWise (CW)"],{"default":"Static"}),
+                "width": ("INT", {"default": 832, "min": 16, "max": MAX_RESOLUTION, "step": 16}),
+                "height": ("INT", {"default": 480, "min": 16, "max": MAX_RESOLUTION, "step": 16}),
+                "length": ("INT", {"default": 81, "min": 1, "max": MAX_RESOLUTION, "step": 4}),
+            },
+            "optional":{
+                "speed":("FLOAT",{"default":1.0, "min": 0, "max": 10.0, "step": 0.1}),
+                "fx":("FLOAT",{"default":0.5, "min": 0, "max": 1, "step": 0.000000001}),
+                "fy":("FLOAT",{"default":0.5, "min": 0, "max": 1, "step": 0.000000001}),
+                "cx":("FLOAT",{"default":0.5, "min": 0, "max": 1, "step": 0.01}),
+                "cy":("FLOAT",{"default":0.5, "min": 0, "max": 1, "step": 0.01}),
+            }
+
+        }
+
+    RETURN_TYPES = ("WAN_CAMERA_EMBEDDING","INT","INT","INT")
+    RETURN_NAMES = ("camera_embedding","width","height","length")
+    FUNCTION = "run"
+    CATEGORY = "camera"
+
+    def run(self, camera_pose, width, height, length, speed=1.0,  fx=0.5, fy=0.5, cx=0.5, cy=0.5):
+        """
+        Use Camera trajectory as extrinsic parameters to calculate Plücker embeddings (Sitzmannet al., 2021)
+        Adapted from https://github.com/aigc-apps/VideoX-Fun/blob/main/comfyui/comfyui_nodes.py
+        """
+        motion_list = [camera_pose]
+        speed = speed
+        angle = np.array(CAMERA_DICT[motion_list[0]]["angle"])
+        T = np.array(CAMERA_DICT[motion_list[0]]["T"])
+        RT = get_camera_motion(angle, T, speed, length)
+
+        trajs=[]
+        for cp in RT.tolist():
+            traj=[fx,fy,cx,cy,0,0]
+            traj.extend(cp[0])
+            traj.extend(cp[1])
+            traj.extend(cp[2])
+            traj.extend([0,0,0,1])
+            trajs.append(traj)
+
+        cam_params = np.array([[float(x) for x in pose] for pose in trajs])
+        cam_params = np.concatenate([np.zeros_like(cam_params[:, :1]), cam_params], 1)
+        control_camera_video = process_pose_params(cam_params, width=width, height=height)
+        control_camera_video = control_camera_video.permute([3, 0, 1, 2]).unsqueeze(0).to(device=comfy.model_management.intermediate_device())
+
+        control_camera_video = torch.concat(
+            [
+                torch.repeat_interleave(control_camera_video[:, :, 0:1], repeats=4, dim=2),
+                control_camera_video[:, :, 1:]
+            ], dim=2
+        ).transpose(1, 2)
+
+        # Reshape, transpose, and view into desired shape
+        b, f, c, h, w = control_camera_video.shape
+        control_camera_video = control_camera_video.contiguous().view(b, f // 4, 4, c, h, w).transpose(2, 3)
+        control_camera_video = control_camera_video.contiguous().view(b, f // 4, c * 4, h, w).transpose(1, 2)
+
+        return (control_camera_video, width, height, length)
+
+
+NODE_CLASS_MAPPINGS = {
+    "WanCameraEmbedding": WanCameraEmbedding,
+}

comfy_extras/nodes_cond.py

@@ -31,6 +31,7 @@ class T5TokenizerOptions:
             }
         }
 
+    CATEGORY = "_for_testing/conditioning"
     RETURN_TYPES = ("CLIP",)
     FUNCTION = "set_options"
 

comfy_extras/nodes_hunyuan.py

@@ -77,7 +77,7 @@ class HunyuanImageToVideo:
                              "height": ("INT", {"default": 480, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 16}),
                              "length": ("INT", {"default": 53, "min": 1, "max": nodes.MAX_RESOLUTION, "step": 4}),
                              "batch_size": ("INT", {"default": 1, "min": 1, "max": 4096}),
-                             "guidance_type": (["v1 (concat)", "v2 (replace)"], )
+                             "guidance_type": (["v1 (concat)", "v2 (replace)", "custom"], )
                 },
                 "optional": {"start_image": ("IMAGE", ),
                 }}
@@ -101,10 +101,12 @@ class HunyuanImageToVideo:
 
             if guidance_type == "v1 (concat)":
                 cond = {"concat_latent_image": concat_latent_image, "concat_mask": mask}
-            else:
+            elif guidance_type == "v2 (replace)":
                 cond = {'guiding_frame_index': 0}
                 latent[:, :, :concat_latent_image.shape[2]] = concat_latent_image
                 out_latent["noise_mask"] = mask
+            elif guidance_type == "custom":
+                cond = {"ref_latent": concat_latent_image}
 
             positive = node_helpers.conditioning_set_values(positive, cond)
 

comfy_extras/nodes_images.py

@@ -13,6 +13,7 @@ import os
 import re
 from io import BytesIO
 from inspect import cleandoc
+import torch
 
 from comfy.comfy_types import FileLocator
 
@@ -74,6 +75,24 @@ class ImageFromBatch:
         s = s_in[batch_index:batch_index + length].clone()
         return (s,)
 
+
+class ImageAddNoise:
+    @classmethod
+    def INPUT_TYPES(s):
+        return {"required": { "image": ("IMAGE",),
+                              "seed": ("INT", {"default": 0, "min": 0, "max": 0xffffffffffffffff, "control_after_generate": True, "tooltip": "The random seed used for creating the noise."}),
+                              "strength": ("FLOAT", {"default": 0.5, "min": 0.0, "max": 1.0, "step": 0.01}),
+                              }}
+    RETURN_TYPES = ("IMAGE",)
+    FUNCTION = "repeat"
+
+    CATEGORY = "image"
+
+    def repeat(self, image, seed, strength):
+        generator = torch.manual_seed(seed)
+        s = torch.clip((image + strength * torch.randn(image.size(), generator=generator, device="cpu").to(image)), min=0.0, max=1.0)
+        return (s,)
+
 class SaveAnimatedWEBP:
     def __init__(self):
         self.output_dir = folder_paths.get_output_directory()
@@ -295,6 +314,7 @@ NODE_CLASS_MAPPINGS = {
     "ImageCrop": ImageCrop,
     "RepeatImageBatch": RepeatImageBatch,
     "ImageFromBatch": ImageFromBatch,
+    "ImageAddNoise": ImageAddNoise,
     "SaveAnimatedWEBP": SaveAnimatedWEBP,
     "SaveAnimatedPNG": SaveAnimatedPNG,
     "SaveSVGNode": SaveSVGNode,

comfy_extras/nodes_load_3d.py

@@ -2,6 +2,10 @@ import nodes
 import folder_paths
 import os
 
+from comfy.comfy_types import IO
+from comfy_api.input_impl import VideoFromFile
+
+
 def normalize_path(path):
     return path.replace('\\', '/')
 
@@ -21,8 +25,8 @@ class Load3D():
             "height": ("INT", {"default": 1024, "min": 1, "max": 4096, "step": 1}),
         }}
 
-    RETURN_TYPES = ("IMAGE", "MASK", "STRING", "IMAGE", "IMAGE", "LOAD3D_CAMERA")
-    RETURN_NAMES = ("image", "mask", "mesh_path", "normal", "lineart", "camera_info")
+    RETURN_TYPES = ("IMAGE", "MASK", "STRING", "IMAGE", "IMAGE", "LOAD3D_CAMERA", IO.VIDEO)
+    RETURN_NAMES = ("image", "mask", "mesh_path", "normal", "lineart", "camera_info", "recording_video")
 
     FUNCTION = "process"
     EXPERIMENTAL = True
@@ -41,7 +45,14 @@ class Load3D():
         normal_image, ignore_mask2 = load_image_node.load_image(image=normal_path)
         lineart_image, ignore_mask3 = load_image_node.load_image(image=lineart_path)
 
-        return output_image, output_mask, model_file, normal_image, lineart_image, image['camera_info']
+        video = None
+
+        if image['recording'] != "":
+            recording_video_path = folder_paths.get_annotated_filepath(image['recording'])
+
+            video = VideoFromFile(recording_video_path)
+
+        return output_image, output_mask, model_file, normal_image, lineart_image, image['camera_info'], video
 
 class Load3DAnimation():
     @classmethod
@@ -59,8 +70,8 @@ class Load3DAnimation():
             "height": ("INT", {"default": 1024, "min": 1, "max": 4096, "step": 1}),
         }}
 
-    RETURN_TYPES = ("IMAGE", "MASK", "STRING", "IMAGE", "LOAD3D_CAMERA")
-    RETURN_NAMES = ("image", "mask", "mesh_path", "normal", "camera_info")
+    RETURN_TYPES = ("IMAGE", "MASK", "STRING", "IMAGE", "LOAD3D_CAMERA", IO.VIDEO)
+    RETURN_NAMES = ("image", "mask", "mesh_path", "normal", "camera_info", "recording_video")
 
     FUNCTION = "process"
     EXPERIMENTAL = True
@@ -77,7 +88,14 @@ class Load3DAnimation():
         ignore_image, output_mask = load_image_node.load_image(image=mask_path)
         normal_image, ignore_mask2 = load_image_node.load_image(image=normal_path)
 
-        return output_image, output_mask, model_file, normal_image, image['camera_info']
+        video = None
+
+        if image['recording'] != "":
+            recording_video_path = folder_paths.get_annotated_filepath(image['recording'])
+
+            video = VideoFromFile(recording_video_path)
+
+        return output_image, output_mask, model_file, normal_image, image['camera_info'], video
 
 class Preview3D():
     @classmethod

comfy_extras/nodes_string.py

@@ -0,0 +1,323 @@
+import re
+
+from comfy.comfy_types.node_typing import IO
+
+class StringConcatenate():
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "string_a": (IO.STRING, {"multiline": True}),
+                "string_b": (IO.STRING, {"multiline": True}),
+                "delimiter": (IO.STRING, {"multiline": False, "default": ""})
+            }
+        }
+
+    RETURN_TYPES = (IO.STRING,)
+    FUNCTION = "execute"
+    CATEGORY = "utils/string"
+
+    def execute(self, string_a, string_b, delimiter, **kwargs):
+        return delimiter.join((string_a, string_b)),
+
+class StringSubstring():
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "string": (IO.STRING, {"multiline": True}),
+                "start": (IO.INT, {}),
+                "end": (IO.INT, {}),
+            }
+        }
+
+    RETURN_TYPES = (IO.STRING,)
+    FUNCTION = "execute"
+    CATEGORY = "utils/string"
+
+    def execute(self, string, start, end, **kwargs):
+        return string[start:end],
+
+class StringLength():
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "string": (IO.STRING, {"multiline": True})
+            }
+        }
+
+    RETURN_TYPES = (IO.INT,)
+    RETURN_NAMES = ("length",)
+    FUNCTION = "execute"
+    CATEGORY = "utils/string"
+
+    def execute(self, string, **kwargs):
+        length = len(string)
+
+        return length,
+
+class CaseConverter():
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "string": (IO.STRING, {"multiline": True}),
+                "mode": (IO.COMBO, {"options": ["UPPERCASE", "lowercase", "Capitalize", "Title Case"]})
+            }
+        }
+
+    RETURN_TYPES = (IO.STRING,)
+    FUNCTION = "execute"
+    CATEGORY = "utils/string"
+
+    def execute(self, string, mode, **kwargs):
+        if mode == "UPPERCASE":
+            result = string.upper()
+        elif mode == "lowercase":
+            result = string.lower()
+        elif mode == "Capitalize":
+            result = string.capitalize()
+        elif mode == "Title Case":
+            result = string.title()
+        else:
+            result = string
+
+        return result,
+
+
+class StringTrim():
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "string": (IO.STRING, {"multiline": True}),
+                "mode": (IO.COMBO, {"options": ["Both", "Left", "Right"]})
+            }
+        }
+
+    RETURN_TYPES = (IO.STRING,)
+    FUNCTION = "execute"
+    CATEGORY = "utils/string"
+
+    def execute(self, string, mode, **kwargs):
+        if mode == "Both":
+            result = string.strip()
+        elif mode == "Left":
+            result = string.lstrip()
+        elif mode == "Right":
+            result = string.rstrip()
+        else:
+            result = string
+
+        return result,
+
+class StringReplace():
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "string": (IO.STRING, {"multiline": True}),
+                "find": (IO.STRING, {"multiline": True}),
+                "replace": (IO.STRING, {"multiline": True})
+            }
+        }
+
+    RETURN_TYPES = (IO.STRING,)
+    FUNCTION = "execute"
+    CATEGORY = "utils/string"
+
+    def execute(self, string, find, replace, **kwargs):
+        result = string.replace(find, replace)
+        return result,
+
+
+class StringContains():
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "string": (IO.STRING, {"multiline": True}),
+                "substring": (IO.STRING, {"multiline": True}),
+                "case_sensitive": (IO.BOOLEAN, {"default": True})
+            }
+        }
+
+    RETURN_TYPES = (IO.BOOLEAN,)
+    RETURN_NAMES = ("contains",)
+    FUNCTION = "execute"
+    CATEGORY = "utils/string"
+
+    def execute(self, string, substring, case_sensitive, **kwargs):
+        if case_sensitive:
+            contains = substring in string
+        else:
+            contains = substring.lower() in string.lower()
+
+        return contains,
+
+
+class StringCompare():
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "string_a": (IO.STRING, {"multiline": True}),
+                "string_b": (IO.STRING, {"multiline": True}),
+                "mode": (IO.COMBO, {"options": ["Starts With", "Ends With", "Equal"]}),
+                "case_sensitive": (IO.BOOLEAN, {"default": True})
+            }
+        }
+
+    RETURN_TYPES = (IO.BOOLEAN,)
+    FUNCTION = "execute"
+    CATEGORY = "utils/string"
+
+    def execute(self, string_a, string_b, mode, case_sensitive, **kwargs):
+        if case_sensitive:
+            a = string_a
+            b = string_b
+        else:
+            a = string_a.lower()
+            b = string_b.lower()
+
+        if mode == "Equal":
+            return a == b,
+        elif mode == "Starts With":
+            return a.startswith(b),
+        elif mode == "Ends With":
+            return a.endswith(b),
+
+class RegexMatch():
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "string": (IO.STRING, {"multiline": True}),
+                "regex_pattern": (IO.STRING, {"multiline": True}),
+                "case_insensitive": (IO.BOOLEAN, {"default": True}),
+                "multiline": (IO.BOOLEAN, {"default": False}),
+                "dotall": (IO.BOOLEAN, {"default": False})
+            }
+        }
+
+    RETURN_TYPES = (IO.BOOLEAN,)
+    RETURN_NAMES = ("matches",)
+    FUNCTION = "execute"
+    CATEGORY = "utils/string"
+
+    def execute(self, string, regex_pattern, case_insensitive, multiline, dotall, **kwargs):
+        flags = 0
+
+        if case_insensitive:
+            flags |= re.IGNORECASE
+        if multiline:
+            flags |= re.MULTILINE
+        if dotall:
+            flags |= re.DOTALL
+
+        try:
+            match = re.search(regex_pattern, string, flags)
+            result = match is not None
+
+        except re.error:
+            result = False
+
+        return result,
+
+
+class RegexExtract():
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "string": (IO.STRING, {"multiline": True}),
+                "regex_pattern": (IO.STRING, {"multiline": True}),
+                "mode": (IO.COMBO, {"options": ["First Match", "All Matches", "First Group", "All Groups"]}),
+                "case_insensitive": (IO.BOOLEAN, {"default": True}),
+                "multiline": (IO.BOOLEAN, {"default": False}),
+                "dotall": (IO.BOOLEAN, {"default": False}),
+                "group_index": (IO.INT, {"default": 1, "min": 0, "max": 100})
+            }
+        }
+
+    RETURN_TYPES = (IO.STRING,)
+    FUNCTION = "execute"
+    CATEGORY = "utils/string"
+
+    def execute(self, string, regex_pattern, mode, case_insensitive, multiline, dotall, group_index, **kwargs):
+        join_delimiter = "\n"
+
+        flags = 0
+        if case_insensitive:
+            flags |= re.IGNORECASE
+        if multiline:
+            flags |= re.MULTILINE
+        if dotall:
+            flags |= re.DOTALL
+
+        try:
+            if mode == "First Match":
+                match = re.search(regex_pattern, string, flags)
+                if match:
+                    result = match.group(0)
+                else:
+                    result = ""
+
+            elif mode == "All Matches":
+                matches = re.findall(regex_pattern, string, flags)
+                if matches:
+                    if isinstance(matches[0], tuple):
+                        result = join_delimiter.join([m[0] for m in matches])
+                    else:
+                        result = join_delimiter.join(matches)
+                else:
+                    result = ""
+
+            elif mode == "First Group":
+                match = re.search(regex_pattern, string, flags)
+                if match and len(match.groups()) >= group_index:
+                    result = match.group(group_index)
+                else:
+                    result = ""
+
+            elif mode == "All Groups":
+                matches = re.finditer(regex_pattern, string, flags)
+                results = []
+                for match in matches:
+                    if match.groups() and len(match.groups()) >= group_index:
+                        results.append(match.group(group_index))
+                result = join_delimiter.join(results)
+            else:
+                result = ""
+
+        except re.error:
+            result = ""
+
+        return result,
+
+NODE_CLASS_MAPPINGS = {
+    "StringConcatenate": StringConcatenate,
+    "StringSubstring": StringSubstring,
+    "StringLength": StringLength,
+    "CaseConverter": CaseConverter,
+    "StringTrim": StringTrim,
+    "StringReplace": StringReplace,
+    "StringContains": StringContains,
+    "StringCompare": StringCompare,
+    "RegexMatch": RegexMatch,
+    "RegexExtract": RegexExtract
+}
+
+NODE_DISPLAY_NAME_MAPPINGS = {
+    "StringConcatenate": "Concatenate",
+    "StringSubstring": "Substring",
+    "StringLength": "Length",
+    "CaseConverter": "Case Converter",
+    "StringTrim": "Trim",
+    "StringReplace": "Replace",
+    "StringContains": "Contains",
+    "StringCompare": "Compare",
+    "RegexMatch": "Regex Match",
+    "RegexExtract": "Regex Extract"
+}

comfy_extras/nodes_wan.py

@@ -297,6 +297,52 @@ class TrimVideoLatent:
         samples_out["samples"] = s1[:, :, trim_amount:]
         return (samples_out,)
 
+class WanCameraImageToVideo:
+    @classmethod
+    def INPUT_TYPES(s):
+        return {"required": {"positive": ("CONDITIONING", ),
+                             "negative": ("CONDITIONING", ),
+                             "vae": ("VAE", ),
+                             "width": ("INT", {"default": 832, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 16}),
+                             "height": ("INT", {"default": 480, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 16}),
+                             "length": ("INT", {"default": 81, "min": 1, "max": nodes.MAX_RESOLUTION, "step": 4}),
+                             "batch_size": ("INT", {"default": 1, "min": 1, "max": 4096}),
+                },
+                "optional": {"clip_vision_output": ("CLIP_VISION_OUTPUT", ),
+                             "start_image": ("IMAGE", ),
+                             "camera_conditions": ("WAN_CAMERA_EMBEDDING", ),
+                }}
+
+    RETURN_TYPES = ("CONDITIONING", "CONDITIONING", "LATENT")
+    RETURN_NAMES = ("positive", "negative", "latent")
+    FUNCTION = "encode"
+
+    CATEGORY = "conditioning/video_models"
+
+    def encode(self, positive, negative, vae, width, height, length, batch_size, start_image=None, clip_vision_output=None, camera_conditions=None):
+        latent = torch.zeros([batch_size, 16, ((length - 1) // 4) + 1, height // 8, width // 8], device=comfy.model_management.intermediate_device())
+        concat_latent = torch.zeros([batch_size, 16, ((length - 1) // 4) + 1, height // 8, width // 8], device=comfy.model_management.intermediate_device())
+        concat_latent = comfy.latent_formats.Wan21().process_out(concat_latent)
+
+        if start_image is not None:
+            start_image = comfy.utils.common_upscale(start_image[:length].movedim(-1, 1), width, height, "bilinear", "center").movedim(1, -1)
+            concat_latent_image = vae.encode(start_image[:, :, :, :3])
+            concat_latent[:,:,:concat_latent_image.shape[2]] = concat_latent_image[:,:,:concat_latent.shape[2]]
+
+            positive = node_helpers.conditioning_set_values(positive, {"concat_latent_image": concat_latent})
+            negative = node_helpers.conditioning_set_values(negative, {"concat_latent_image": concat_latent})
+
+        if camera_conditions is not None:
+            positive = node_helpers.conditioning_set_values(positive, {'camera_conditions': camera_conditions})
+            negative = node_helpers.conditioning_set_values(negative, {'camera_conditions': camera_conditions})
+
+        if clip_vision_output is not None:
+            positive = node_helpers.conditioning_set_values(positive, {"clip_vision_output": clip_vision_output})
+            negative = node_helpers.conditioning_set_values(negative, {"clip_vision_output": clip_vision_output})
+
+        out_latent = {}
+        out_latent["samples"] = latent
+        return (positive, negative, out_latent)
 
 NODE_CLASS_MAPPINGS = {
     "WanImageToVideo": WanImageToVideo,
@@ -305,4 +351,5 @@ NODE_CLASS_MAPPINGS = {
     "WanFirstLastFrameToVideo": WanFirstLastFrameToVideo,
     "WanVaceToVideo": WanVaceToVideo,
     "TrimVideoLatent": TrimVideoLatent,
+    "WanCameraImageToVideo": WanCameraImageToVideo,
 }

fix_torch.py

@@ -1,28 +0,0 @@
-import importlib.util
-import shutil
-import os
-import ctypes
-import logging
-
-
-def fix_pytorch_libomp():
-    """
-    Fix PyTorch libomp DLL issue on Windows by copying the correct DLL file if needed.
-    """
-    torch_spec = importlib.util.find_spec("torch")
-    for folder in torch_spec.submodule_search_locations:
-        lib_folder = os.path.join(folder, "lib")
-        test_file = os.path.join(lib_folder, "fbgemm.dll")
-        dest = os.path.join(lib_folder, "libomp140.x86_64.dll")
-        if os.path.exists(dest):
-            break
-
-        with open(test_file, "rb") as f:
-            contents = f.read()
-            if b"libomp140.x86_64.dll" not in contents:
-                break
-        try:
-            ctypes.cdll.LoadLibrary(test_file)
-        except FileNotFoundError:
-            logging.warning("Detected pytorch version with libomp issue, patching.")
-            shutil.copyfile(os.path.join(lib_folder, "libiomp5md.dll"), dest)

main.py

@@ -125,13 +125,6 @@ if __name__ == "__main__":
 
     import cuda_malloc
 
-if args.windows_standalone_build:
-    try:
-        from fix_torch import fix_pytorch_libomp
-        fix_pytorch_libomp()
-    except:
-        pass
-
 import comfy.utils
 
 import execution

nodes.py

@@ -1940,7 +1940,7 @@ class ImagePadForOutpaint:
 
         mask[top:top + d2, left:left + d3] = t
 
-        return (new_image, mask)
+        return (new_image, mask.unsqueeze(0))
 
 
 NODE_CLASS_MAPPINGS = {
@@ -2261,8 +2261,11 @@ def init_builtin_extra_nodes():
         "nodes_optimalsteps.py",
         "nodes_hidream.py",
         "nodes_fresca.py",
+        "nodes_apg.py",
         "nodes_preview_any.py",
         "nodes_ace.py",
+        "nodes_string.py",
+        "nodes_camera_trajectory.py",
     ]
 
     import_failed = []

requirements.txt

@@ -1,4 +1,4 @@
-comfyui-frontend-package==1.18.10
+comfyui-frontend-package==1.19.9
 comfyui-workflow-templates==0.1.14
 torch
 torchsde

script_examples/basic_api_example.py

@@ -101,6 +101,14 @@ prompt_text = """
 
 def queue_prompt(prompt):
     p = {"prompt": prompt}
+
+    # If the workflow contains API nodes, you can add a Comfy API key to the `extra_data`` field of the payload.
+    # p["extra_data"] = {
+    #     "api_key_comfy_org": "comfyui-87d01e28d*******************************************************"  # replace with real key
+    # }
+    # See: https://docs.comfy.org/tutorials/api-nodes/overview
+    # Generate a key here: https://platform.comfy.org/login
+
     data = json.dumps(p).encode('utf-8')
     req =  request.Request("http://127.0.0.1:8188/prompt", data=data)
     request.urlopen(req)

tests-api/README.md

@@ -0,0 +1,74 @@
+# ComfyUI API Testing
+
+This directory contains tests for validating the ComfyUI OpenAPI specification against a running instance of ComfyUI.
+
+## Setup
+
+1. Install the required dependencies:
+
+```bash
+pip install -r requirements.txt
+```
+
+2. Make sure you have a running instance of ComfyUI (default: http://127.0.0.1:8188)
+
+## Running the Tests
+
+Run all tests with pytest:
+
+```bash
+cd tests-api
+pytest
+```
+
+Run specific test files:
+
+```bash
+pytest test_spec_validation.py
+pytest test_endpoint_existence.py
+pytest test_schema_validation.py
+pytest test_api_by_tag.py
+```
+
+Run tests with more verbose output:
+
+```bash
+pytest -v
+```
+
+## Test Categories
+
+The tests are organized into several categories:
+
+1. **Spec Validation**: Validates that the OpenAPI specification is valid.
+2. **Endpoint Existence**: Tests that the endpoints defined in the spec exist on the server.
+3. **Schema Validation**: Tests that the server responses match the schemas defined in the spec.
+4. **Tag-Based Tests**: Tests that the API's tag organization is consistent.
+
+## Using a Different Server
+
+By default, the tests connect to `http://127.0.0.1:8188`. To test against a different server, set the `COMFYUI_SERVER_URL` environment variable:
+
+```bash
+COMFYUI_SERVER_URL=http://example.com:8188 pytest
+```
+
+## Test Structure
+
+- `conftest.py`: Contains pytest fixtures used by the tests.
+- `utils/`: Contains utility functions for working with the OpenAPI spec.
+- `test_*.py`: The actual test files.
+- `resources/`: Contains resources used by the tests (e.g., sample workflows).
+
+## Extending the Tests
+
+To add new tests:
+
+1. For testing new endpoints, add them to the appropriate test file based on their category.
+2. For testing more complex functionality, create a new test file following the established patterns.
+
+## Notes
+
+- Tests that require a running server will be skipped if the server is not available.
+- Some tests may fail if the server doesn't match the specification exactly.
+- The tests don't modify any data on the server (they're read-only).

tests-api/conftest.py

@@ -0,0 +1,142 @@
+"""
+Test fixtures for API testing
+"""
+import os
+import pytest
+import yaml
+import requests
+import logging
+from typing import Dict, Any, Generator, Optional
+from urllib.parse import urljoin
+
+# Setup logging
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(__name__)
+
+# Default server configuration
+DEFAULT_SERVER_URL = "http://127.0.0.1:8188"
+
+
+@pytest.fixture(scope="session")
+def api_spec_path() -> str:
+    """
+    Get the path to the OpenAPI specification file
+
+    Returns:
+        Path to the OpenAPI specification file
+    """
+    return os.path.abspath(os.path.join(
+        os.path.dirname(__file__),
+        "..",
+        "openapi.yaml"
+    ))
+
+
+@pytest.fixture(scope="session")
+def api_spec(api_spec_path: str) -> Dict[str, Any]:
+    """
+    Load the OpenAPI specification
+
+    Args:
+        api_spec_path: Path to the spec file
+
+    Returns:
+        Parsed OpenAPI specification
+    """
+    with open(api_spec_path, 'r') as f:
+        return yaml.safe_load(f)
+
+
+@pytest.fixture(scope="session")
+def base_url() -> str:
+    """
+    Get the base URL for the API server
+
+    Returns:
+        Base URL string
+    """
+    # Allow overriding via environment variable
+    return os.environ.get("COMFYUI_SERVER_URL", DEFAULT_SERVER_URL)
+
+
+@pytest.fixture(scope="session")
+def server_available(base_url: str) -> bool:
+    """
+    Check if the server is available
+
+    Args:
+        base_url: Base URL for the API
+
+    Returns:
+        True if the server is available, False otherwise
+    """
+    try:
+        response = requests.get(base_url, timeout=2)
+        return response.status_code == 200
+    except requests.RequestException:
+        logger.warning(f"Server at {base_url} is not available")
+        return False
+
+
+@pytest.fixture
+def api_client(base_url: str) -> Generator[Optional[requests.Session], None, None]:
+    """
+    Create a requests session for API testing
+
+    Args:
+        base_url: Base URL for the API
+
+    Yields:
+        Requests session configured for the API
+    """
+    session = requests.Session()
+
+    # Helper function to construct URLs
+    def get_url(path: str) -> str:
+        # Paths in the OpenAPI spec already include /api prefix where needed
+        return urljoin(base_url, path)
+
+    # Add url helper to the session
+    session.get_url = get_url  # type: ignore
+
+    yield session
+
+    # Cleanup
+    session.close()
+
+
+@pytest.fixture
+def api_get_json(api_client: requests.Session):
+    """
+    Helper fixture for making GET requests and parsing JSON responses
+
+    Args:
+        api_client: API client session
+
+    Returns:
+        Function that makes GET requests and returns JSON
+    """
+    def _get_json(path: str, **kwargs):
+        url = api_client.get_url(path)  # type: ignore
+        response = api_client.get(url, **kwargs)
+
+        if response.status_code == 200:
+            try:
+                return response.json()
+            except ValueError:
+                return None
+        return None
+
+    return _get_json
+
+
+@pytest.fixture
+def require_server(server_available):
+    """
+    Skip tests if server is not available
+
+    Args:
+        server_available: Whether the server is available
+    """
+    if not server_available:
+        pytest.skip("Server is not available")

tests-api/requirements.txt

@@ -0,0 +1,6 @@
+pytest>=7.0.0
+pytest-asyncio>=0.21.0
+openapi-spec-validator>=0.5.0
+jsonschema>=4.17.0
+requests>=2.28.0
+pyyaml>=6.0.0

tests-api/test_api_by_tag.py

@@ -0,0 +1,279 @@
+"""
+Tests for API endpoints grouped by tags
+"""
+import pytest
+import logging
+import sys
+import os
+from typing import Dict, Any, Set
+
+# Use a direct import with the full path
+current_dir = os.path.dirname(os.path.abspath(__file__))
+sys.path.insert(0, current_dir)
+
+# Define functions inline to avoid import issues
+def get_all_endpoints(spec):
+    """
+    Extract all endpoints from an OpenAPI spec
+    """
+    endpoints = []
+
+    for path, path_item in spec['paths'].items():
+        for method, operation in path_item.items():
+            if method.lower() not in ['get', 'post', 'put', 'delete', 'patch']:
+                continue
+
+            endpoints.append({
+                'path': path,
+                'method': method.lower(),
+                'tags': operation.get('tags', []),
+                'operation_id': operation.get('operationId', ''),
+                'summary': operation.get('summary', '')
+            })
+
+    return endpoints
+
+def get_all_tags(spec):
+    """
+    Get all tags used in the API spec
+    """
+    tags = set()
+
+    for path_item in spec['paths'].values():
+        for operation in path_item.values():
+            if isinstance(operation, dict) and 'tags' in operation:
+                tags.update(operation['tags'])
+
+    return tags
+
+def extract_endpoints_by_tag(spec, tag):
+    """
+    Extract all endpoints with a specific tag
+    """
+    endpoints = []
+
+    for path, path_item in spec['paths'].items():
+        for method, operation in path_item.items():
+            if method.lower() not in ['get', 'post', 'put', 'delete', 'patch']:
+                continue
+
+            if tag in operation.get('tags', []):
+                endpoints.append({
+                    'path': path,
+                    'method': method.lower(),
+                    'operation_id': operation.get('operationId', ''),
+                    'summary': operation.get('summary', '')
+                })
+
+    return endpoints
+
+# Setup logging
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(__name__)
+
+
+@pytest.fixture
+def api_tags(api_spec: Dict[str, Any]) -> Set[str]:
+    """
+    Get all tags from the API spec
+
+    Args:
+        api_spec: Loaded OpenAPI spec
+
+    Returns:
+        Set of tag names
+    """
+    return get_all_tags(api_spec)
+
+
+def test_api_has_tags(api_tags: Set[str]):
+    """
+    Test that the API has defined tags
+
+    Args:
+        api_tags: Set of tags
+    """
+    assert len(api_tags) > 0, "API spec should have at least one tag"
+
+    # Log the tags
+    logger.info(f"API spec has the following tags: {sorted(api_tags)}")
+
+
+@pytest.mark.parametrize("tag", [
+    "workflow",
+    "image",
+    "model",
+    "node",
+    "system"
+])
+def test_core_tags_exist(api_tags: Set[str], tag: str):
+    """
+    Test that core tags exist in the API spec
+
+    Args:
+        api_tags: Set of tags
+        tag: Tag to check
+    """
+    assert tag in api_tags, f"API spec should have '{tag}' tag"
+
+
+def test_workflow_tag_has_endpoints(api_spec: Dict[str, Any]):
+    """
+    Test that the 'workflow' tag has appropriate endpoints
+
+    Args:
+        api_spec: Loaded OpenAPI spec
+    """
+    endpoints = extract_endpoints_by_tag(api_spec, "workflow")
+
+    assert len(endpoints) > 0, "No endpoints found with 'workflow' tag"
+
+    # Check for key workflow endpoints
+    endpoint_paths = [e["path"] for e in endpoints]
+    assert "/prompt" in endpoint_paths, "Workflow tag should include /prompt endpoint"
+
+    # Log the endpoints
+    logger.info(f"Found {len(endpoints)} endpoints with 'workflow' tag:")
+    for e in endpoints:
+        logger.info(f"  {e['method'].upper()} {e['path']}")
+
+
+def test_image_tag_has_endpoints(api_spec: Dict[str, Any]):
+    """
+    Test that the 'image' tag has appropriate endpoints
+
+    Args:
+        api_spec: Loaded OpenAPI spec
+    """
+    endpoints = extract_endpoints_by_tag(api_spec, "image")
+
+    assert len(endpoints) > 0, "No endpoints found with 'image' tag"
+
+    # Check for key image endpoints
+    endpoint_paths = [e["path"] for e in endpoints]
+    assert "/upload/image" in endpoint_paths, "Image tag should include /upload/image endpoint"
+    assert "/view" in endpoint_paths, "Image tag should include /view endpoint"
+
+    # Log the endpoints
+    logger.info(f"Found {len(endpoints)} endpoints with 'image' tag:")
+    for e in endpoints:
+        logger.info(f"  {e['method'].upper()} {e['path']}")
+
+
+def test_model_tag_has_endpoints(api_spec: Dict[str, Any]):
+    """
+    Test that the 'model' tag has appropriate endpoints
+
+    Args:
+        api_spec: Loaded OpenAPI spec
+    """
+    endpoints = extract_endpoints_by_tag(api_spec, "model")
+
+    assert len(endpoints) > 0, "No endpoints found with 'model' tag"
+
+    # Check for key model endpoints
+    endpoint_paths = [e["path"] for e in endpoints]
+    assert "/models" in endpoint_paths, "Model tag should include /models endpoint"
+
+    # Log the endpoints
+    logger.info(f"Found {len(endpoints)} endpoints with 'model' tag:")
+    for e in endpoints:
+        logger.info(f"  {e['method'].upper()} {e['path']}")
+
+
+def test_node_tag_has_endpoints(api_spec: Dict[str, Any]):
+    """
+    Test that the 'node' tag has appropriate endpoints
+
+    Args:
+        api_spec: Loaded OpenAPI spec
+    """
+    endpoints = extract_endpoints_by_tag(api_spec, "node")
+
+    assert len(endpoints) > 0, "No endpoints found with 'node' tag"
+
+    # Check for key node endpoints
+    endpoint_paths = [e["path"] for e in endpoints]
+    assert "/object_info" in endpoint_paths, "Node tag should include /object_info endpoint"
+
+    # Log the endpoints
+    logger.info(f"Found {len(endpoints)} endpoints with 'node' tag:")
+    for e in endpoints:
+        logger.info(f"  {e['method'].upper()} {e['path']}")
+
+
+def test_system_tag_has_endpoints(api_spec: Dict[str, Any]):
+    """
+    Test that the 'system' tag has appropriate endpoints
+
+    Args:
+        api_spec: Loaded OpenAPI spec
+    """
+    endpoints = extract_endpoints_by_tag(api_spec, "system")
+
+    assert len(endpoints) > 0, "No endpoints found with 'system' tag"
+
+    # Check for key system endpoints
+    endpoint_paths = [e["path"] for e in endpoints]
+    assert "/system_stats" in endpoint_paths, "System tag should include /system_stats endpoint"
+
+    # Log the endpoints
+    logger.info(f"Found {len(endpoints)} endpoints with 'system' tag:")
+    for e in endpoints:
+        logger.info(f"  {e['method'].upper()} {e['path']}")
+
+
+def test_internal_tag_has_endpoints(api_spec: Dict[str, Any]):
+    """
+    Test that the 'internal' tag has appropriate endpoints
+
+    Args:
+        api_spec: Loaded OpenAPI spec
+    """
+    endpoints = extract_endpoints_by_tag(api_spec, "internal")
+
+    assert len(endpoints) > 0, "No endpoints found with 'internal' tag"
+
+    # Check for key internal endpoints
+    endpoint_paths = [e["path"] for e in endpoints]
+    assert "/internal/logs" in endpoint_paths, "Internal tag should include /internal/logs endpoint"
+
+    # Log the endpoints
+    logger.info(f"Found {len(endpoints)} endpoints with 'internal' tag:")
+    for e in endpoints:
+        logger.info(f"  {e['method'].upper()} {e['path']}")
+
+
+def test_operation_ids_match_tag(api_spec: Dict[str, Any]):
+    """
+    Test that operation IDs follow a consistent pattern with their tag
+
+    Args:
+        api_spec: Loaded OpenAPI spec
+    """
+    failures = []
+
+    for path, path_item in api_spec['paths'].items():
+        for method, operation in path_item.items():
+            if method in ['get', 'post', 'put', 'delete', 'patch']:
+                if 'operationId' in operation and 'tags' in operation and operation['tags']:
+                    op_id = operation['operationId']
+                    primary_tag = operation['tags'][0].lower()
+
+                    # Check if operationId starts with primary tag prefix
+                    # This is a common convention, but might need adjusting
+                    if not (op_id.startswith(primary_tag) or
+                            any(op_id.lower().startswith(f"{tag.lower()}") for tag in operation['tags'])):
+                        failures.append({
+                            'path': path,
+                            'method': method,
+                            'operationId': op_id,
+                            'primary_tag': primary_tag
+                        })
+
+    # Log failures for diagnosis but don't fail the test
+    # as this is a style/convention check
+    if failures:
+        logger.warning(f"Found {len(failures)} operationIds that don't align with their tags:")
+        for f in failures:
+            logger.warning(f"  {f['method'].upper()} {f['path']} - operationId: {f['operationId']}, primary tag: {f['primary_tag']}")

tests-api/test_endpoint_existence.py

@@ -0,0 +1,243 @@
+"""
+Tests for endpoint existence and basic response codes
+"""
+import pytest
+import requests
+import logging
+import sys
+import os
+from typing import Dict, Any, List
+from urllib.parse import urljoin
+
+# Use a direct import with the full path
+current_dir = os.path.dirname(os.path.abspath(__file__))
+sys.path.insert(0, current_dir)
+
+# Define get_all_endpoints function inline to avoid import issues
+def get_all_endpoints(spec):
+    """
+    Extract all endpoints from an OpenAPI spec
+
+    Args:
+        spec: Parsed OpenAPI specification
+
+    Returns:
+        List of dicts with path, method, and tags for each endpoint
+    """
+    endpoints = []
+
+    for path, path_item in spec['paths'].items():
+        for method, operation in path_item.items():
+            if method.lower() not in ['get', 'post', 'put', 'delete', 'patch']:
+                continue
+
+            endpoints.append({
+                'path': path,
+                'method': method.lower(),
+                'tags': operation.get('tags', []),
+                'operation_id': operation.get('operationId', ''),
+                'summary': operation.get('summary', '')
+            })
+
+    return endpoints
+
+# Setup logging
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(__name__)
+
+
+@pytest.fixture
+def all_endpoints(api_spec: Dict[str, Any]) -> List[Dict[str, Any]]:
+    """
+    Get all endpoints from the API spec
+
+    Args:
+        api_spec: Loaded OpenAPI spec
+
+    Returns:
+        List of endpoint information
+    """
+    return get_all_endpoints(api_spec)
+
+
+def test_endpoints_exist(all_endpoints: List[Dict[str, Any]]):
+    """
+    Test that endpoints are defined in the spec
+
+    Args:
+        all_endpoints: List of endpoint information
+    """
+    # Simple check that we have endpoints defined
+    assert len(all_endpoints) > 0, "No endpoints defined in the OpenAPI spec"
+
+    # Log the endpoints for informational purposes
+    logger.info(f"Found {len(all_endpoints)} endpoints in the OpenAPI spec")
+    for endpoint in all_endpoints:
+        logger.info(f"{endpoint['method'].upper()} {endpoint['path']} - {endpoint['summary']}")
+
+
+@pytest.mark.parametrize("endpoint_path", [
+    "/",                  # Root path (doesn't have /api prefix)
+    "/api/prompt",        # Get prompt info
+    "/api/queue",         # Get queue
+    "/api/models",        # Get model types
+    "/api/object_info",   # Get node info
+    "/api/system_stats"   # Get system stats
+])
+def test_basic_get_endpoints(require_server, api_client, endpoint_path: str):
+    """
+    Test that basic GET endpoints exist and respond
+
+    Args:
+        require_server: Fixture that skips if server is not available
+        api_client: API client fixture
+        endpoint_path: Path to test
+    """
+    url = api_client.get_url(endpoint_path)  # type: ignore
+
+    try:
+        response = api_client.get(url)
+
+        # We're just checking that the endpoint exists and returns some kind of response
+        # Not necessarily a 200 status code
+        assert response.status_code not in [404, 405], f"Endpoint {endpoint_path} does not exist"
+
+        logger.info(f"Endpoint {endpoint_path} exists with status code {response.status_code}")
+
+    except requests.RequestException as e:
+        pytest.fail(f"Request to {endpoint_path} failed: {str(e)}")
+
+
+def test_websocket_endpoint_exists(require_server, base_url: str):
+    """
+    Test that the WebSocket endpoint exists
+
+    Args:
+        require_server: Fixture that skips if server is not available
+        base_url: Base server URL
+    """
+    # WebSocket endpoint path from OpenAPI spec
+    ws_url = urljoin(base_url, "/api/ws")
+
+    # For WebSocket, we can't use a normal GET request
+    # Instead, we make a HEAD request to check if the endpoint exists
+    try:
+        response = requests.head(ws_url)
+
+        # WebSocket endpoints often return a 400 Bad Request for HEAD requests
+        # but a 404 would indicate the endpoint doesn't exist
+        assert response.status_code != 404, "WebSocket endpoint /ws does not exist"
+
+        logger.info(f"WebSocket endpoint exists with status code {response.status_code}")
+
+    except requests.RequestException as e:
+        pytest.fail(f"Request to WebSocket endpoint failed: {str(e)}")
+
+
+def test_api_models_folder_endpoint(require_server, api_client):
+    """
+    Test that the /models/{folder} endpoint exists and responds
+
+    Args:
+        require_server: Fixture that skips if server is not available
+        api_client: API client fixture
+    """
+    # First get available model types
+    models_url = api_client.get_url("/api/models")  # type: ignore
+
+    try:
+        models_response = api_client.get(models_url)
+        assert models_response.status_code == 200, "Failed to get model types"
+
+        model_types = models_response.json()
+
+        # Skip if no model types available
+        if not model_types:
+            pytest.skip("No model types available to test")
+
+        # Test with the first model type
+        model_type = model_types[0]
+        models_folder_url = api_client.get_url(f"/api/models/{model_type}")  # type: ignore
+
+        folder_response = api_client.get(models_folder_url)
+
+        # We're just checking that the endpoint exists
+        assert folder_response.status_code != 404, f"Endpoint /api/models/{model_type} does not exist"
+
+        logger.info(f"Endpoint /api/models/{model_type} exists with status code {folder_response.status_code}")
+
+    except requests.RequestException as e:
+        pytest.fail(f"Request failed: {str(e)}")
+    except (ValueError, KeyError, IndexError) as e:
+        pytest.fail(f"Failed to process response: {str(e)}")
+
+
+def test_api_object_info_node_endpoint(require_server, api_client):
+    """
+    Test that the /object_info/{node_class} endpoint exists and responds
+
+    Args:
+        require_server: Fixture that skips if server is not available
+        api_client: API client fixture
+    """
+    # First get available node classes
+    objects_url = api_client.get_url("/api/object_info")  # type: ignore
+
+    try:
+        objects_response = api_client.get(objects_url)
+        assert objects_response.status_code == 200, "Failed to get object info"
+
+        node_classes = objects_response.json()
+
+        # Skip if no node classes available
+        if not node_classes:
+            pytest.skip("No node classes available to test")
+
+        # Test with the first node class
+        node_class = next(iter(node_classes.keys()))
+        node_url = api_client.get_url(f"/api/object_info/{node_class}")  # type: ignore
+
+        node_response = api_client.get(node_url)
+
+        # We're just checking that the endpoint exists
+        assert node_response.status_code != 404, f"Endpoint /api/object_info/{node_class} does not exist"
+
+        logger.info(f"Endpoint /api/object_info/{node_class} exists with status code {node_response.status_code}")
+
+    except requests.RequestException as e:
+        pytest.fail(f"Request failed: {str(e)}")
+    except (ValueError, KeyError, StopIteration) as e:
+        pytest.fail(f"Failed to process response: {str(e)}")
+
+
+def test_internal_endpoints_exist(require_server, api_client, base_url: str):
+    """
+    Test that internal endpoints exist
+
+    Args:
+        require_server: Fixture that skips if server is not available
+        api_client: API client fixture
+        base_url: Base server URL
+    """
+    internal_endpoints = [
+        "/internal/logs",
+        "/internal/logs/raw",
+        "/internal/folder_paths",
+        "/internal/files/output"
+    ]
+
+    for endpoint in internal_endpoints:
+        # Internal endpoints don't use the /api/ prefix
+        url = urljoin(base_url, endpoint)
+
+        try:
+            response = requests.get(url)
+
+            # We're just checking that the endpoint exists
+            assert response.status_code != 404, f"Endpoint {endpoint} does not exist"
+
+            logger.info(f"Endpoint {endpoint} exists with status code {response.status_code}")
+
+        except requests.RequestException as e:
+            logger.warning(f"Request to {endpoint} failed: {str(e)}")
+            # Don't fail the test as internal endpoints might be restricted

tests-api/test_schema_validation.py

@@ -0,0 +1,453 @@
+"""
+Tests for validating API responses against OpenAPI schema
+"""
+import pytest
+import requests
+import logging
+import sys
+import os
+import json
+from typing import Dict, Any
+
+# Use a direct import with the full path
+current_dir = os.path.dirname(os.path.abspath(__file__))
+sys.path.insert(0, current_dir)
+
+# Define validation functions inline to avoid import issues
+def get_endpoint_schema(
+    spec,
+    path,
+    method,
+    status_code = '200'
+):
+    """
+    Extract response schema for a specific endpoint from OpenAPI spec
+    """
+    method = method.lower()
+
+    # Handle path not found
+    if path not in spec['paths']:
+        return None
+
+    # Handle method not found
+    if method not in spec['paths'][path]:
+        return None
+
+    # Handle status code not found
+    responses = spec['paths'][path][method].get('responses', {})
+    if status_code not in responses:
+        return None
+
+    # Handle no content defined
+    if 'content' not in responses[status_code]:
+        return None
+
+    # Get schema from first content type
+    content_types = responses[status_code]['content']
+    first_content_type = next(iter(content_types))
+
+    if 'schema' not in content_types[first_content_type]:
+        return None
+
+    return content_types[first_content_type]['schema']
+
+def resolve_schema_refs(schema, spec):
+    """
+    Resolve $ref references in a schema
+    """
+    if not isinstance(schema, dict):
+        return schema
+
+    result = {}
+
+    for key, value in schema.items():
+        if key == '$ref' and isinstance(value, str) and value.startswith('#/'):
+            # Handle reference
+            ref_path = value[2:].split('/')
+            ref_value = spec
+            for path_part in ref_path:
+                ref_value = ref_value.get(path_part, {})
+
+            # Recursively resolve any refs in the referenced schema
+            ref_value = resolve_schema_refs(ref_value, spec)
+            result.update(ref_value)
+        elif isinstance(value, dict):
+            # Recursively resolve refs in nested dictionaries
+            result[key] = resolve_schema_refs(value, spec)
+        elif isinstance(value, list):
+            # Recursively resolve refs in list items
+            result[key] = [
+                resolve_schema_refs(item, spec) if isinstance(item, dict) else item
+                for item in value
+            ]
+        else:
+            # Pass through other values
+            result[key] = value
+
+    return result
+
+def validate_response(
+    response_data,
+    spec,
+    path,
+    method,
+    status_code = '200'
+):
+    """
+    Validate a response against the OpenAPI schema
+    """
+    schema = get_endpoint_schema(spec, path, method, status_code)
+
+    if schema is None:
+        return {
+            'valid': False,
+            'errors': [f"No schema found for {method.upper()} {path} with status {status_code}"]
+        }
+
+    # Resolve any $ref in the schema
+    resolved_schema = resolve_schema_refs(schema, spec)
+
+    try:
+        import jsonschema
+        jsonschema.validate(instance=response_data, schema=resolved_schema)
+        return {'valid': True, 'errors': []}
+    except jsonschema.exceptions.ValidationError as e:
+        # Extract more detailed error information
+        path = ".".join(str(p) for p in e.path) if e.path else "root"
+        instance = e.instance if not isinstance(e.instance, dict) else "..."
+        schema_path = ".".join(str(p) for p in e.schema_path) if e.schema_path else "unknown"
+
+        detailed_error = (
+            f"Validation error at path: {path}\n"
+            f"Schema path: {schema_path}\n"
+            f"Error message: {e.message}\n"
+            f"Failed instance: {instance}\n"
+        )
+
+        return {'valid': False, 'errors': [detailed_error]}
+
+# Setup logging
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(__name__)
+
+
+@pytest.mark.parametrize("endpoint_path,method", [
+    ("/api/system_stats", "get"),
+    ("/api/prompt", "get"),
+    ("/api/queue", "get"),
+    ("/api/models", "get"),
+    ("/api/embeddings", "get")
+])
+def test_response_schema_validation(
+    require_server,
+    api_client,
+    api_spec: Dict[str, Any],
+    endpoint_path: str,
+    method: str
+):
+    """
+    Test that API responses match the defined schema
+
+    Args:
+        require_server: Fixture that skips if server is not available
+        api_client: API client fixture
+        api_spec: Loaded OpenAPI spec
+        endpoint_path: Path to test
+        method: HTTP method to test
+    """
+    url = api_client.get_url(endpoint_path)  # type: ignore
+
+    # Skip if no schema defined
+    schema = get_endpoint_schema(api_spec, endpoint_path, method)
+    if not schema:
+        pytest.skip(f"No schema defined for {method.upper()} {endpoint_path}")
+
+    try:
+        if method.lower() == "get":
+            response = api_client.get(url)
+        else:
+            pytest.skip(f"Method {method} not implemented for automated testing")
+            return
+
+        # Skip if response is not 200
+        if response.status_code != 200:
+            pytest.skip(f"Endpoint {endpoint_path} returned status {response.status_code}")
+            return
+
+        # Skip if response is not JSON
+        try:
+            response_data = response.json()
+        except ValueError:
+            pytest.skip(f"Endpoint {endpoint_path} did not return valid JSON")
+            return
+
+        # Special handling for system_stats endpoint
+        if endpoint_path == '/api/system_stats' and isinstance(response_data, dict):
+            # Remove null index fields before validation
+            for device in response_data.get('devices', []):
+                if 'index' in device and device['index'] is None:
+                    del device['index']
+
+        # Validate the response
+        validation_result = validate_response(
+            response_data,
+            api_spec,
+            endpoint_path,
+            method
+        )
+
+        if validation_result['valid']:
+            logger.info(f"Response from {method.upper()} {endpoint_path} matches schema")
+        else:
+            for error in validation_result['errors']:
+                logger.error(f"Validation error for {method.upper()} {endpoint_path}: {error}")
+
+        assert validation_result['valid'], f"Response from {method.upper()} {endpoint_path} does not match schema"
+
+    except requests.RequestException as e:
+        pytest.fail(f"Request to {endpoint_path} failed: {str(e)}")
+
+
+def test_system_stats_response(require_server, api_client, api_spec: Dict[str, Any]):
+    """
+    Test the system_stats endpoint response in detail
+
+    Args:
+        require_server: Fixture that skips if server is not available
+        api_client: API client fixture
+        api_spec: Loaded OpenAPI spec
+    """
+    url = api_client.get_url("/api/system_stats")  # type: ignore
+
+    try:
+        response = api_client.get(url)
+
+        assert response.status_code == 200, "Failed to get system stats"
+
+        # Parse response
+        stats = response.json()
+
+        # Validate high-level structure
+        assert 'system' in stats, "Response missing 'system' field"
+        assert 'devices' in stats, "Response missing 'devices' field"
+
+        # Validate system fields
+        system = stats['system']
+        assert 'os' in system, "System missing 'os' field"
+        assert 'ram_total' in system, "System missing 'ram_total' field"
+        assert 'ram_free' in system, "System missing 'ram_free' field"
+        assert 'comfyui_version' in system, "System missing 'comfyui_version' field"
+
+        # Validate devices fields
+        devices = stats['devices']
+        assert isinstance(devices, list), "Devices should be a list"
+
+        if devices:
+            device = devices[0]
+            assert 'name' in device, "Device missing 'name' field"
+            assert 'type' in device, "Device missing 'type' field"
+            assert 'vram_total' in device, "Device missing 'vram_total' field"
+            assert 'vram_free' in device, "Device missing 'vram_free' field"
+
+        # Remove null index fields before validation
+        # This is needed because ComfyUI returns null for CPU device index
+        for device in stats.get('devices', []):
+            if 'index' in device and device['index'] is None:
+                del device['index']
+
+        # Perform schema validation
+        validation_result = validate_response(
+            stats,
+            api_spec,
+            "/api/system_stats",
+            "get"
+        )
+
+        # Print detailed error if validation fails
+        if not validation_result['valid']:
+            for error in validation_result['errors']:
+                logger.error(f"Validation error for /system_stats: {error}")
+
+            # Print schema details for debugging
+            schema = get_endpoint_schema(api_spec, "/system_stats", "get")
+            if schema:
+                logger.error(f"Schema structure:\n{json.dumps(schema, indent=2)}")
+
+            # Print sample of the response
+            logger.error(f"Response:\n{json.dumps(stats, indent=2)}")
+
+        assert validation_result['valid'], "System stats response does not match schema"
+
+    except requests.RequestException as e:
+        pytest.fail(f"Request to /api/system_stats failed: {str(e)}")
+
+
+def test_models_listing_response(require_server, api_client, api_spec: Dict[str, Any]):
+    """
+    Test the models endpoint response
+
+    Args:
+        require_server: Fixture that skips if server is not available
+        api_client: API client fixture
+        api_spec: Loaded OpenAPI spec
+    """
+    url = api_client.get_url("/api/models")  # type: ignore
+
+    try:
+        response = api_client.get(url)
+
+        assert response.status_code == 200, "Failed to get models"
+
+        # Parse response
+        models = response.json()
+
+        # Validate it's a list
+        assert isinstance(models, list), "Models response should be a list"
+
+        # Each item should be a string
+        for model in models:
+            assert isinstance(model, str), "Each model type should be a string"
+
+        # Perform schema validation
+        validation_result = validate_response(
+            models,
+            api_spec,
+            "/api/models",
+            "get"
+        )
+
+        # Print detailed error if validation fails
+        if not validation_result['valid']:
+            for error in validation_result['errors']:
+                logger.error(f"Validation error for /models: {error}")
+
+            # Print schema details for debugging
+            schema = get_endpoint_schema(api_spec, "/models", "get")
+            if schema:
+                logger.error(f"Schema structure:\n{json.dumps(schema, indent=2)}")
+
+            # Print response
+            sample_models = models[:5] if isinstance(models, list) else models
+            logger.error(f"Models response:\n{json.dumps(sample_models, indent=2)}")
+
+        assert validation_result['valid'], "Models response does not match schema"
+
+    except requests.RequestException as e:
+        pytest.fail(f"Request to /api/models failed: {str(e)}")
+
+
+def test_object_info_response(require_server, api_client, api_spec: Dict[str, Any]):
+    """
+    Test the object_info endpoint response
+
+    Args:
+        require_server: Fixture that skips if server is not available
+        api_client: API client fixture
+        api_spec: Loaded OpenAPI spec
+    """
+    url = api_client.get_url("/api/object_info")  # type: ignore
+
+    try:
+        response = api_client.get(url)
+
+        assert response.status_code == 200, "Failed to get object info"
+
+        # Parse response
+        objects = response.json()
+
+        # Validate it's an object
+        assert isinstance(objects, dict), "Object info response should be an object"
+
+        # Check if we have any objects
+        if objects:
+            # Get the first object
+            first_obj_name = next(iter(objects.keys()))
+            first_obj = objects[first_obj_name]
+
+            # Validate first object has required fields
+            assert 'input' in first_obj, "Object missing 'input' field"
+            assert 'output' in first_obj, "Object missing 'output' field"
+            assert 'name' in first_obj, "Object missing 'name' field"
+
+        # Perform schema validation
+        validation_result = validate_response(
+            objects,
+            api_spec,
+            "/api/object_info",
+            "get"
+        )
+
+        # Print detailed error if validation fails
+        if not validation_result['valid']:
+            for error in validation_result['errors']:
+                logger.error(f"Validation error for /object_info: {error}")
+
+            # Print schema details for debugging
+            schema = get_endpoint_schema(api_spec, "/object_info", "get")
+            if schema:
+                logger.error(f"Schema structure:\n{json.dumps(schema, indent=2)}")
+
+            # Also print a small sample of the response
+            sample = dict(list(objects.items())[:1]) if objects else {}
+            logger.error(f"Sample response:\n{json.dumps(sample, indent=2)}")
+
+        assert validation_result['valid'], "Object info response does not match schema"
+
+    except requests.RequestException as e:
+        pytest.fail(f"Request to /api/object_info failed: {str(e)}")
+    except (KeyError, StopIteration) as e:
+        pytest.fail(f"Failed to process response: {str(e)}")
+
+
+def test_queue_response(require_server, api_client, api_spec: Dict[str, Any]):
+    """
+    Test the queue endpoint response
+
+    Args:
+        require_server: Fixture that skips if server is not available
+        api_client: API client fixture
+        api_spec: Loaded OpenAPI spec
+    """
+    url = api_client.get_url("/api/queue")  # type: ignore
+
+    try:
+        response = api_client.get(url)
+
+        assert response.status_code == 200, "Failed to get queue"
+
+        # Parse response
+        queue = response.json()
+
+        # Validate structure
+        assert 'queue_running' in queue, "Queue missing 'queue_running' field"
+        assert 'queue_pending' in queue, "Queue missing 'queue_pending' field"
+
+        # Each should be a list
+        assert isinstance(queue['queue_running'], list), "queue_running should be a list"
+        assert isinstance(queue['queue_pending'], list), "queue_pending should be a list"
+
+        # Perform schema validation
+        validation_result = validate_response(
+            queue,
+            api_spec,
+            "/api/queue",
+            "get"
+        )
+
+        # Print detailed error if validation fails
+        if not validation_result['valid']:
+            for error in validation_result['errors']:
+                logger.error(f"Validation error for /queue: {error}")
+
+            # Print schema details for debugging
+            schema = get_endpoint_schema(api_spec, "/queue", "get")
+            if schema:
+                logger.error(f"Schema structure:\n{json.dumps(schema, indent=2)}")
+
+            # Print response
+            logger.error(f"Queue response:\n{json.dumps(queue, indent=2)}")
+
+        assert validation_result['valid'], "Queue response does not match schema"
+
+    except requests.RequestException as e:
+        pytest.fail(f"Request to /queue failed: {str(e)}")

tests-api/test_spec_validation.py

@@ -0,0 +1,144 @@
+"""
+Tests for validating the OpenAPI specification
+"""
+import pytest
+from openapi_spec_validator import validate_spec
+from openapi_spec_validator.exceptions import OpenAPISpecValidatorError
+from typing import Dict, Any
+
+
+def test_openapi_spec_is_valid(api_spec: Dict[str, Any]):
+    """
+    Test that the OpenAPI specification is valid
+
+    Args:
+        api_spec: Loaded OpenAPI spec
+    """
+    try:
+        validate_spec(api_spec)
+    except OpenAPISpecValidatorError as e:
+        pytest.fail(f"OpenAPI spec validation failed: {str(e)}")
+
+
+def test_spec_has_info(api_spec: Dict[str, Any]):
+    """
+    Test that the OpenAPI spec has the required info section
+
+    Args:
+        api_spec: Loaded OpenAPI spec
+    """
+    assert 'info' in api_spec, "Spec must have info section"
+    assert 'title' in api_spec['info'], "Info must have title"
+    assert 'version' in api_spec['info'], "Info must have version"
+
+
+def test_spec_has_paths(api_spec: Dict[str, Any]):
+    """
+    Test that the OpenAPI spec has paths defined
+
+    Args:
+        api_spec: Loaded OpenAPI spec
+    """
+    assert 'paths' in api_spec, "Spec must have paths section"
+    assert len(api_spec['paths']) > 0, "Spec must have at least one path"
+
+
+def test_spec_has_components(api_spec: Dict[str, Any]):
+    """
+    Test that the OpenAPI spec has components defined
+
+    Args:
+        api_spec: Loaded OpenAPI spec
+    """
+    assert 'components' in api_spec, "Spec must have components section"
+    assert 'schemas' in api_spec['components'], "Components must have schemas"
+
+
+def test_workflow_endpoints_exist(api_spec: Dict[str, Any]):
+    """
+    Test that core workflow endpoints are defined
+
+    Args:
+        api_spec: Loaded OpenAPI spec
+    """
+    assert '/api/prompt' in api_spec['paths'], "Spec must define /api/prompt endpoint"
+    assert 'post' in api_spec['paths']['/api/prompt'], "Spec must define POST /api/prompt"
+    assert 'get' in api_spec['paths']['/api/prompt'], "Spec must define GET /api/prompt"
+
+
+def test_image_endpoints_exist(api_spec: Dict[str, Any]):
+    """
+    Test that core image endpoints are defined
+
+    Args:
+        api_spec: Loaded OpenAPI spec
+    """
+    assert '/api/upload/image' in api_spec['paths'], "Spec must define /api/upload/image endpoint"
+    assert '/api/view' in api_spec['paths'], "Spec must define /api/view endpoint"
+
+
+def test_model_endpoints_exist(api_spec: Dict[str, Any]):
+    """
+    Test that core model endpoints are defined
+
+    Args:
+        api_spec: Loaded OpenAPI spec
+    """
+    assert '/api/models' in api_spec['paths'], "Spec must define /api/models endpoint"
+    assert '/api/models/{folder}' in api_spec['paths'], "Spec must define /api/models/{folder} endpoint"
+
+
+def test_operation_ids_are_unique(api_spec: Dict[str, Any]):
+    """
+    Test that all operationIds are unique
+
+    Args:
+        api_spec: Loaded OpenAPI spec
+    """
+    operation_ids = []
+
+    for path, path_item in api_spec['paths'].items():
+        for method, operation in path_item.items():
+            if method in ['get', 'post', 'put', 'delete', 'patch']:
+                if 'operationId' in operation:
+                    operation_ids.append(operation['operationId'])
+
+    # Check for duplicates
+    duplicates = set([op_id for op_id in operation_ids if operation_ids.count(op_id) > 1])
+    assert len(duplicates) == 0, f"Found duplicate operationIds: {duplicates}"
+
+
+def test_all_endpoints_have_operation_ids(api_spec: Dict[str, Any]):
+    """
+    Test that all endpoints have operationIds
+
+    Args:
+        api_spec: Loaded OpenAPI spec
+    """
+    missing = []
+
+    for path, path_item in api_spec['paths'].items():
+        for method, operation in path_item.items():
+            if method in ['get', 'post', 'put', 'delete', 'patch']:
+                if 'operationId' not in operation:
+                    missing.append(f"{method.upper()} {path}")
+
+    assert len(missing) == 0, f"Found endpoints without operationIds: {missing}"
+
+
+def test_all_endpoints_have_tags(api_spec: Dict[str, Any]):
+    """
+    Test that all endpoints have tags
+
+    Args:
+        api_spec: Loaded OpenAPI spec
+    """
+    missing = []
+
+    for path, path_item in api_spec['paths'].items():
+        for method, operation in path_item.items():
+            if method in ['get', 'post', 'put', 'delete', 'patch']:
+                if 'tags' not in operation or not operation['tags']:
+                    missing.append(f"{method.upper()} {path}")
+
+    assert len(missing) == 0, f"Found endpoints without tags: {missing}"

tests-api/utils/schema_utils.py

@@ -0,0 +1,157 @@
+"""
+Utilities for working with OpenAPI schemas
+"""
+from typing import Any, Dict, List, Optional, Set, Tuple
+
+
+def extract_required_parameters(
+    spec: Dict[str, Any],
+    path: str,
+    method: str
+) -> Tuple[List[Dict[str, Any]], List[Dict[str, Any]]]:
+    """
+    Extract required parameters for a specific endpoint
+
+    Args:
+        spec: Parsed OpenAPI specification
+        path: API path (e.g., '/prompt')
+        method: HTTP method (e.g., 'get', 'post')
+
+    Returns:
+        Tuple of (path_params, query_params) containing required parameters
+    """
+    method = method.lower()
+    path_params = []
+    query_params = []
+
+    # Handle path not found
+    if path not in spec['paths']:
+        return path_params, query_params
+
+    # Handle method not found
+    if method not in spec['paths'][path]:
+        return path_params, query_params
+
+    # Get parameters
+    params = spec['paths'][path][method].get('parameters', [])
+
+    for param in params:
+        if param.get('required', False):
+            if param.get('in') == 'path':
+                path_params.append(param)
+            elif param.get('in') == 'query':
+                query_params.append(param)
+
+    return path_params, query_params
+
+
+def get_request_body_schema(
+    spec: Dict[str, Any],
+    path: str,
+    method: str
+) -> Optional[Dict[str, Any]]:
+    """
+    Get request body schema for a specific endpoint
+
+    Args:
+        spec: Parsed OpenAPI specification
+        path: API path (e.g., '/prompt')
+        method: HTTP method (e.g., 'get', 'post')
+
+    Returns:
+        Request body schema or None if not found
+    """
+    method = method.lower()
+
+    # Handle path not found
+    if path not in spec['paths']:
+        return None
+
+    # Handle method not found
+    if method not in spec['paths'][path]:
+        return None
+
+    # Handle no request body
+    request_body = spec['paths'][path][method].get('requestBody', {})
+    if not request_body or 'content' not in request_body:
+        return None
+
+    # Get schema from first content type
+    content_types = request_body['content']
+    first_content_type = next(iter(content_types))
+
+    if 'schema' not in content_types[first_content_type]:
+        return None
+
+    return content_types[first_content_type]['schema']
+
+
+def extract_endpoints_by_tag(spec: Dict[str, Any], tag: str) -> List[Dict[str, Any]]:
+    """
+    Extract all endpoints with a specific tag
+
+    Args:
+        spec: Parsed OpenAPI specification
+        tag: Tag to filter by
+
+    Returns:
+        List of endpoint details
+    """
+    endpoints = []
+
+    for path, path_item in spec['paths'].items():
+        for method, operation in path_item.items():
+            if method.lower() not in ['get', 'post', 'put', 'delete', 'patch']:
+                continue
+
+            if tag in operation.get('tags', []):
+                endpoints.append({
+                    'path': path,
+                    'method': method.lower(),
+                    'operation_id': operation.get('operationId', ''),
+                    'summary': operation.get('summary', '')
+                })
+
+    return endpoints
+
+
+def get_all_tags(spec: Dict[str, Any]) -> Set[str]:
+    """
+    Get all tags used in the API spec
+
+    Args:
+        spec: Parsed OpenAPI specification
+
+    Returns:
+        Set of tag names
+    """
+    tags = set()
+
+    for path_item in spec['paths'].values():
+        for operation in path_item.values():
+            if isinstance(operation, dict) and 'tags' in operation:
+                tags.update(operation['tags'])
+
+    return tags
+
+
+def get_schema_examples(spec: Dict[str, Any]) -> Dict[str, Any]:
+    """
+    Extract all examples from component schemas
+
+    Args:
+        spec: Parsed OpenAPI specification
+
+    Returns:
+        Dict mapping schema names to examples
+    """
+    examples = {}
+
+    if 'components' not in spec or 'schemas' not in spec['components']:
+        return examples
+
+    for name, schema in spec['components']['schemas'].items():
+        if 'example' in schema:
+            examples[name] = schema['example']
+
+    return examples

tests-api/utils/validation.py

@@ -0,0 +1,201 @@
+"""
+Utilities for API response validation against OpenAPI spec
+"""
+import yaml
+import jsonschema
+from typing import Any, Dict, List, Optional, Union
+
+
+def load_openapi_spec(spec_path: str) -> Dict[str, Any]:
+    """
+    Load the OpenAPI specification from a YAML file
+
+    Args:
+        spec_path: Path to the OpenAPI specification file
+
+    Returns:
+        Dict containing the parsed OpenAPI spec
+    """
+    with open(spec_path, 'r') as f:
+        return yaml.safe_load(f)
+
+
+def get_endpoint_schema(
+    spec: Dict[str, Any],
+    path: str,
+    method: str,
+    status_code: str = '200'
+) -> Optional[Dict[str, Any]]:
+    """
+    Extract response schema for a specific endpoint from OpenAPI spec
+
+    Args:
+        spec: Parsed OpenAPI specification
+        path: API path (e.g., '/prompt')
+        method: HTTP method (e.g., 'get', 'post')
+        status_code: HTTP status code to get schema for
+
+    Returns:
+        Schema dict or None if not found
+    """
+    method = method.lower()
+
+    # Handle path not found
+    if path not in spec['paths']:
+        return None
+
+    # Handle method not found
+    if method not in spec['paths'][path]:
+        return None
+
+    # Handle status code not found
+    responses = spec['paths'][path][method].get('responses', {})
+    if status_code not in responses:
+        return None
+
+    # Handle no content defined
+    if 'content' not in responses[status_code]:
+        return None
+
+    # Get schema from first content type
+    content_types = responses[status_code]['content']
+    first_content_type = next(iter(content_types))
+
+    if 'schema' not in content_types[first_content_type]:
+        return None
+
+    return content_types[first_content_type]['schema']
+
+
+def resolve_schema_refs(schema: Dict[str, Any], spec: Dict[str, Any]) -> Dict[str, Any]:
+    """
+    Resolve $ref references in a schema and convert OpenAPI nullable to JSON Schema
+
+    Args:
+        schema: Schema that may contain references
+        spec: Full OpenAPI spec with component definitions
+
+    Returns:
+        Schema with references resolved
+    """
+    if not isinstance(schema, dict):
+        return schema
+
+    result = {}
+
+    # Check if this schema has nullable: true with a type
+    if schema.get('nullable') is True and 'type' in schema:
+        # Convert OpenAPI nullable syntax to JSON Schema oneOf
+        original_type = schema['type']
+        result['oneOf'] = [
+            {'type': original_type},
+            {'type': 'null'}
+        ]
+        # Copy other properties except nullable and type
+        for key, value in schema.items():
+            if key not in ['nullable', 'type']:
+                if isinstance(value, dict):
+                    result[key] = resolve_schema_refs(value, spec)
+                elif isinstance(value, list):
+                    result[key] = [
+                        resolve_schema_refs(item, spec) if isinstance(item, dict) else item
+                        for item in value
+                    ]
+                else:
+                    result[key] = value
+    else:
+        # Normal processing
+        for key, value in schema.items():
+            if key == '$ref' and isinstance(value, str) and value.startswith('#/'):
+                # Handle reference
+                ref_path = value[2:].split('/')
+                ref_value = spec
+                for path_part in ref_path:
+                    ref_value = ref_value.get(path_part, {})
+
+                # Recursively resolve any refs in the referenced schema
+                ref_value = resolve_schema_refs(ref_value, spec)
+                result.update(ref_value)
+            elif isinstance(value, dict):
+                # Recursively resolve refs in nested dictionaries
+                result[key] = resolve_schema_refs(value, spec)
+            elif isinstance(value, list):
+                # Recursively resolve refs in list items
+                result[key] = [
+                    resolve_schema_refs(item, spec) if isinstance(item, dict) else item
+                    for item in value
+                ]
+            else:
+                # Pass through other values (skip nullable as it's OpenAPI specific)
+                if key != 'nullable':
+                    result[key] = value
+
+    return result
+
+
+def validate_response(
+    response_data: Union[Dict[str, Any], List[Any]],
+    spec: Dict[str, Any],
+    path: str,
+    method: str,
+    status_code: str = '200'
+) -> Dict[str, Any]:
+    """
+    Validate a response against the OpenAPI schema
+
+    Args:
+        response_data: Response data to validate
+        spec: Parsed OpenAPI specification
+        path: API path (e.g., '/prompt')
+        method: HTTP method (e.g., 'get', 'post')
+        status_code: HTTP status code to validate against
+
+    Returns:
+        Dict with validation result containing:
+            - valid: bool indicating if validation passed
+            - errors: List of validation errors if any
+    """
+    schema = get_endpoint_schema(spec, path, method, status_code)
+
+    if schema is None:
+        return {
+            'valid': False,
+            'errors': [f"No schema found for {method.upper()} {path} with status {status_code}"]
+        }
+
+    # Resolve any $ref in the schema
+    resolved_schema = resolve_schema_refs(schema, spec)
+
+    try:
+        jsonschema.validate(instance=response_data, schema=resolved_schema)
+        return {'valid': True, 'errors': []}
+    except jsonschema.exceptions.ValidationError as e:
+        return {'valid': False, 'errors': [str(e)]}
+
+
+def get_all_endpoints(spec: Dict[str, Any]) -> List[Dict[str, Any]]:
+    """
+    Extract all endpoints from an OpenAPI spec
+
+    Args:
+        spec: Parsed OpenAPI specification
+
+    Returns:
+        List of dicts with path, method, and tags for each endpoint
+    """
+    endpoints = []
+
+    for path, path_item in spec['paths'].items():
+        for method, operation in path_item.items():
+            if method.lower() not in ['get', 'post', 'put', 'delete', 'patch']:
+                continue
+
+            endpoints.append({
+                'path': path,
+                'method': method.lower(),
+                'tags': operation.get('tags', []),
+                'operation_id': operation.get('operationId', ''),
+                'summary': operation.get('summary', '')
+            })
+
+    return endpoints

tests-unit/comfy_api_test/video_types_test.py

@@ -0,0 +1,239 @@
+import pytest
+import torch
+import tempfile
+import os
+import av
+import io
+from fractions import Fraction
+from comfy_api.input_impl.video_types import VideoFromFile, VideoFromComponents
+from comfy_api.util.video_types import VideoComponents
+from comfy_api.input.basic_types import AudioInput
+from av.error import InvalidDataError
+
+EPSILON = 0.0001
+
+
+@pytest.fixture
+def sample_images():
+    """3-frame 2x2 RGB video tensor"""
+    return torch.rand(3, 2, 2, 3)
+
+
+@pytest.fixture
+def sample_audio():
+    """Stereo audio with 44.1kHz sample rate"""
+    return AudioInput(
+        {
+            "waveform": torch.rand(1, 2, 1000),
+            "sample_rate": 44100,
+        }
+    )
+
+
+@pytest.fixture
+def video_components(sample_images, sample_audio):
+    """VideoComponents with images, audio, and metadata"""
+    return VideoComponents(
+        images=sample_images,
+        audio=sample_audio,
+        frame_rate=Fraction(30),
+        metadata={"test": "metadata"},
+    )
+
+
+def create_test_video(width=4, height=4, frames=3, fps=30):
+    """Helper to create a temporary video file"""
+    tmp = tempfile.NamedTemporaryFile(suffix=".mp4", delete=False)
+    with av.open(tmp.name, mode="w") as container:
+        stream = container.add_stream("h264", rate=fps)
+        stream.width = width
+        stream.height = height
+        stream.pix_fmt = "yuv420p"
+
+        for i in range(frames):
+            frame = av.VideoFrame.from_ndarray(
+                torch.ones(height, width, 3, dtype=torch.uint8).numpy() * (i * 85),
+                format="rgb24",
+            )
+            frame = frame.reformat(format="yuv420p")
+            packet = stream.encode(frame)
+            container.mux(packet)
+
+        # Flush
+        packet = stream.encode(None)
+        container.mux(packet)
+
+    return tmp.name
+
+
+@pytest.fixture
+def simple_video_file():
+    """4x4 video with 3 frames at 30fps"""
+    file_path = create_test_video()
+    yield file_path
+    os.unlink(file_path)
+
+
+def test_video_from_components_get_duration(video_components):
+    """Duration calculated correctly from frame count and frame rate"""
+    video = VideoFromComponents(video_components)
+    duration = video.get_duration()
+
+    expected_duration = 3.0 / 30.0
+    assert duration == pytest.approx(expected_duration)
+
+
+def test_video_from_components_get_duration_different_frame_rates(sample_images):
+    """Duration correct for different frame rates including fractional"""
+    # Test with 60 fps
+    components_60fps = VideoComponents(images=sample_images, frame_rate=Fraction(60))
+    video_60fps = VideoFromComponents(components_60fps)
+    assert video_60fps.get_duration() == pytest.approx(3.0 / 60.0)
+
+    # Test with fractional frame rate (23.976fps)
+    components_frac = VideoComponents(
+        images=sample_images, frame_rate=Fraction(24000, 1001)
+    )
+    video_frac = VideoFromComponents(components_frac)
+    expected_frac = 3.0 / (24000.0 / 1001.0)
+    assert video_frac.get_duration() == pytest.approx(expected_frac)
+
+
+def test_video_from_components_get_duration_empty_video():
+    """Duration is zero for empty video"""
+    empty_components = VideoComponents(
+        images=torch.zeros(0, 2, 2, 3), frame_rate=Fraction(30)
+    )
+    video = VideoFromComponents(empty_components)
+    assert video.get_duration() == 0.0
+
+
+def test_video_from_components_get_dimensions(video_components):
+    """Dimensions returned correctly from image tensor shape"""
+    video = VideoFromComponents(video_components)
+    width, height = video.get_dimensions()
+    assert width == 2
+    assert height == 2
+
+
+def test_video_from_file_get_duration(simple_video_file):
+    """Duration extracted from file metadata"""
+    video = VideoFromFile(simple_video_file)
+    duration = video.get_duration()
+    assert duration == pytest.approx(0.1, abs=0.01)
+
+
+def test_video_from_file_get_dimensions(simple_video_file):
+    """Dimensions read from stream without decoding frames"""
+    video = VideoFromFile(simple_video_file)
+    width, height = video.get_dimensions()
+    assert width == 4
+    assert height == 4
+
+
+def test_video_from_file_bytesio_input():
+    """VideoFromFile works with BytesIO input"""
+    buffer = io.BytesIO()
+    with av.open(buffer, mode="w", format="mp4") as container:
+        stream = container.add_stream("h264", rate=30)
+        stream.width = 2
+        stream.height = 2
+        stream.pix_fmt = "yuv420p"
+
+        frame = av.VideoFrame.from_ndarray(
+            torch.zeros(2, 2, 3, dtype=torch.uint8).numpy(), format="rgb24"
+        )
+        frame = frame.reformat(format="yuv420p")
+        packet = stream.encode(frame)
+        container.mux(packet)
+        packet = stream.encode(None)
+        container.mux(packet)
+
+    buffer.seek(0)
+    video = VideoFromFile(buffer)
+
+    assert video.get_dimensions() == (2, 2)
+    assert video.get_duration() == pytest.approx(1 / 30, abs=0.01)
+
+
+def test_video_from_file_invalid_file_error():
+    """InvalidDataError raised for non-video files"""
+    with tempfile.NamedTemporaryFile(suffix=".txt", delete=False) as tmp:
+        tmp.write(b"not a video file")
+        tmp.flush()
+        tmp_name = tmp.name
+
+    try:
+        with pytest.raises(InvalidDataError):
+            video = VideoFromFile(tmp_name)
+            video.get_dimensions()
+    finally:
+        os.unlink(tmp_name)
+
+
+def test_video_from_file_audio_only_error():
+    """ValueError raised for audio-only files"""
+    with tempfile.NamedTemporaryFile(suffix=".m4a", delete=False) as tmp:
+        tmp_name = tmp.name
+
+    try:
+        with av.open(tmp_name, mode="w") as container:
+            stream = container.add_stream("aac", rate=44100)
+            stream.sample_rate = 44100
+            stream.format = "fltp"
+
+            audio_data = torch.zeros(1, 1024).numpy()
+            audio_frame = av.AudioFrame.from_ndarray(
+                audio_data, format="fltp", layout="mono"
+            )
+            audio_frame.sample_rate = 44100
+            audio_frame.pts = 0
+            packet = stream.encode(audio_frame)
+            container.mux(packet)
+
+            for packet in stream.encode(None):
+                container.mux(packet)
+
+        with pytest.raises(ValueError, match="No video stream found"):
+            video = VideoFromFile(tmp_name)
+            video.get_dimensions()
+    finally:
+        os.unlink(tmp_name)
+
+
+def test_single_frame_video():
+    """Single frame video has correct duration"""
+    components = VideoComponents(
+        images=torch.rand(1, 10, 10, 3), frame_rate=Fraction(1)
+    )
+    video = VideoFromComponents(components)
+    assert video.get_duration() == 1.0
+
+
+@pytest.mark.parametrize(
+    "frame_rate,expected_fps",
+    [
+        (Fraction(24000, 1001), 24000 / 1001),
+        (Fraction(30000, 1001), 30000 / 1001),
+        (Fraction(25, 1), 25.0),
+        (Fraction(50, 2), 25.0),
+    ],
+)
+def test_fractional_frame_rates(frame_rate, expected_fps):
+    """Duration calculated correctly for various fractional frame rates"""
+    components = VideoComponents(images=torch.rand(100, 4, 4, 3), frame_rate=frame_rate)
+    video = VideoFromComponents(components)
+    duration = video.get_duration()
+    expected_duration = 100.0 / expected_fps
+    assert duration == pytest.approx(expected_duration)
+
+
+def test_duration_consistency(video_components):
+    """get_duration() consistent with manual calculation from components"""
+    video = VideoFromComponents(video_components)
+
+    duration = video.get_duration()
+    components = video.get_components()
+    manual_duration = float(components.images.shape[0] / components.frame_rate)
+
+    assert duration == pytest.approx(manual_duration)