Feat: Support siliconflow.com (#13308)

### What problem does this PR solve? Feat: Support siliconflow.com ### Type of change - [x] New Feature (non-breaking change which adds functionality)
2026-04-30 07:17:47 +08:00 · 2026-03-02 15:37:42 +08:00
parent 1db221f19e
commit 5fc3bd38b0
20 changed files with 530 additions and 17 deletions
--- a/api/apps/llm_app.py
+++ b/api/apps/llm_app.py
@ -34,7 +34,7 @@ from rag.llm import EmbeddingModel, ChatModel, RerankModel, CvModel, TTSModel, O
 def factories():
    try:
        fac = get_allowed_llm_factories()
-        fac = [f.to_dict() for f in fac if f.name not in ["Youdao", "FastEmbed", "BAAI", "Builtin"]]
+        fac = [f.to_dict() for f in fac if f.name not in ["Youdao", "FastEmbed", "BAAI", "Builtin", "siliconflow_intl"]]
        llms = LLMService.get_all()
        mdl_types = {}
        for m in llms:
@ -64,13 +64,22 @@ async def set_api_key():
    # test if api key works
    chat_passed, embd_passed, rerank_passed = False, False, False
    factory = req["llm_factory"]
+    base_url = req.get("base_url", "")
+    source_factory = req.get("source_fid", factory)
    extra = {"provider": factory}
    timeout_seconds = int(os.environ.get("LLM_TIMEOUT_SECONDS", 10))
+    source_llms = list(LLMService.query(fid=source_factory))
+    if not source_llms:
+        msg = f"No models configured for {factory} (source: {source_factory})."
+        if req.get("verify", False):
+            return get_json_result(data={"message": msg, "success": False})
+        return get_data_error_result(message=msg)
+
    msg = ""
-    for llm in LLMService.query(fid=factory):
+    for llm in source_llms:
        if not embd_passed and llm.model_type == LLMType.EMBEDDING.value:
            assert factory in EmbeddingModel, f"Embedding model from {factory} is not supported yet."
-            mdl = EmbeddingModel[factory](req["api_key"], llm.llm_name, base_url=req.get("base_url"))
+            mdl = EmbeddingModel[factory](req["api_key"], llm.llm_name, base_url=base_url)
            try:
                arr, tc = await asyncio.wait_for(
                    asyncio.to_thread(mdl.encode, ["Test if the api key is available"]),
@ -83,7 +92,7 @@ async def set_api_key():
                msg += f"\nFail to access embedding model({llm.llm_name}) using this api key." + str(e)
        elif not chat_passed and llm.model_type == LLMType.CHAT.value:
            assert factory in ChatModel, f"Chat model from {factory} is not supported yet."
-            mdl = ChatModel[factory](req["api_key"], llm.llm_name, base_url=req.get("base_url"), **extra)
+            mdl = ChatModel[factory](req["api_key"], llm.llm_name, base_url=base_url, **extra)
            try:
                m, tc = await asyncio.wait_for(
                    mdl.async_chat(
@ -100,7 +109,7 @@ async def set_api_key():
                msg += f"\nFail to access model({llm.fid}/{llm.llm_name}) using this api key." + str(e)
        elif not rerank_passed and llm.model_type == LLMType.RERANK.value:
            assert factory in RerankModel, f"Re-rank model from {factory} is not supported yet."
-            mdl = RerankModel[factory](req["api_key"], llm.llm_name, base_url=req.get("base_url"))
+            mdl = RerankModel[factory](req["api_key"], llm.llm_name, base_url=base_url)
            try:
                arr, tc = await asyncio.wait_for(
                    asyncio.to_thread(mdl.similarity, "What's the weather?", ["Is it sunny today?"]),
@ -122,12 +131,12 @@ async def set_api_key():
    if msg:
        return get_data_error_result(message=msg)

-    llm_config = {"api_key": req["api_key"], "api_base": req.get("base_url", "")}
+    llm_config = {"api_key": req["api_key"], "api_base": base_url}
    for n in ["model_type", "llm_name"]:
        if n in req:
            llm_config[n] = req[n]

-    for llm in LLMService.query(fid=factory):
+    for llm in source_llms:
        llm_config["max_tokens"] = llm.max_tokens
        if not TenantLLMService.filter_update([TenantLLM.tenant_id == current_user.id, TenantLLM.llm_factory == factory, TenantLLM.llm_name == llm.llm_name], llm_config):
            TenantLLMService.save(
--- a/conf/llm_factories.json
+++ b/conf/llm_factories.json
@ -3245,6 +3245,463 @@
                }
            ]
        },
+        {
+            "name": "siliconflow_intl",
+            "logo": "",
+            "tags": "LLM,TEXT EMBEDDING,TEXT RE-RANK,IMAGE2TEXT,TTS",
+            "status": "1",
+            "rank": "781",
+            "llm": [
+                {
+                    "llm_name": "meta-llama/Meta-Llama-3.1-8B-Instruct",
+                    "tags": "LLM,CHAT,33k",
+                    "max_tokens": 33000,
+                    "model_type": "chat",
+                    "is_tools": false
+                },
+                {
+                    "llm_name": "MiniMaxAI/MiniMax-M2.5",
+                    "tags": "LLM,CHAT,197k",
+                    "max_tokens": 197000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "zai-org/GLM-5",
+                    "tags": "LLM,CHAT,205k",
+                    "max_tokens": 205000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "stepfun-ai/Step-3.5-Flash",
+                    "tags": "LLM,CHAT,262k",
+                    "max_tokens": 262000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "moonshotai/Kimi-K2.5",
+                    "tags": "LLM,CHAT,262k",
+                    "max_tokens": 262000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "MiniMaxAI/MiniMax-M2.1",
+                    "tags": "LLM,CHAT,197k",
+                    "max_tokens": 197000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "zai-org/GLM-4.7",
+                    "tags": "LLM,CHAT,205k",
+                    "max_tokens": 205000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "deepseek-ai/DeepSeek-V3.2",
+                    "tags": "LLM,CHAT,164k",
+                    "max_tokens": 164000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "deepseek-ai/DeepSeek-V3.2-Exp",
+                    "tags": "LLM,CHAT,164k",
+                    "max_tokens": 164000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "zai-org/GLM-4.6V",
+                    "tags": "LLM,CHAT,131k",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "deepseek-ai/DeepSeek-V3.1-Terminus",
+                    "tags": "LLM,CHAT,164k",
+                    "max_tokens": 164000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "deepseek-ai/DeepSeek-V3.1",
+                    "tags": "LLM,CHAT,164k",
+                    "max_tokens": 164000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "deepseek-ai/DeepSeek-V3",
+                    "tags": "LLM,CHAT,164k",
+                    "max_tokens": 164000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "deepseek-ai/DeepSeek-R1",
+                    "tags": "LLM,CHAT,154k",
+                    "max_tokens": 154000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "nex-agi/DeepSeek-V3.1-Nex-N1",
+                    "tags": "LLM,CHAT,164k",
+                    "max_tokens": 164000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen3-VL-32B-Instruct",
+                    "tags": "LLM,CHAT,262k",
+                    "max_tokens": 262000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen3-VL-32B-Thinking",
+                    "tags": "LLM,CHAT,262k",
+                    "max_tokens": 262000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "zai-org/GLM-4.5V",
+                    "tags": "LLM,CHAT,66k",
+                    "max_tokens": 66000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "inclusionAI/Ling-mini-2.0",
+                    "tags": "LLM,CHAT,131k",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "inclusionAI/Ring-flash-2.0",
+                    "tags": "LLM,CHAT,131k",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": false
+                },
+                {
+                    "llm_name": "inclusionAI/Ling-flash-2.0",
+                    "tags": "LLM,CHAT,131k",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "tencent/Hunyuan-MT-7B",
+                    "tags": "LLM,CHAT,32k",
+                    "max_tokens": 32000,
+                    "model_type": "chat",
+                    "is_tools": false
+                },
+                {
+                    "llm_name": "Qwen/Qwen3-Omni-30B-A3B-Captioner",
+                    "tags": "LLM,CHAT,131k",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen3-Omni-30B-A3B-Thinking",
+                    "tags": "LLM,CHAT,131k",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen3-Omni-30B-A3B-Instruct",
+                    "tags": "LLM,CHAT,65k",
+                    "max_tokens": 65000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen3-Next-80B-A3B-Thinking",
+                    "tags": "LLM,CHAT,262k",
+                    "max_tokens": 262000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen3-Next-80B-A3B-Instruct",
+                    "tags": "LLM,CHAT,262k",
+                    "max_tokens": 262000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen3-Coder-480B-A35B-Instruct",
+                    "tags": "LLM,CHAT,262k",
+                    "max_tokens": 262000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen3-Coder-30B-A3B-Instruct",
+                    "tags": "LLM,CHAT,262k",
+                    "max_tokens": 262000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen3-30B-A3B-Thinking-2507",
+                    "tags": "LLM,CHAT,262k",
+                    "max_tokens": 262000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen3-30B-A3B-Instruct-2507",
+                    "tags": "LLM,CHAT,262k",
+                    "max_tokens": 262000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen3-235B-A22B-Instruct-2507",
+                    "tags": "LLM,CHAT,262k",
+                    "max_tokens": 262000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen3-235B-A22B-Thinking-2507",
+                    "tags": "LLM,CHAT,262k",
+                    "max_tokens": 262000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "ByteDance-Seed/Seed-OSS-36B-Instruct",
+                    "tags": "LLM,CHAT,262k",
+                    "max_tokens": 262000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "baidu/ERNIE-4.5-300B-A47B",
+                    "tags": "LLM,CHAT,131k",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": false
+                },
+                {
+                    "llm_name": "tencent/Hunyuan-A13B-Instruct",
+                    "tags": "LLM,CHAT,131k",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": false
+                },
+                {
+                    "llm_name": "moonshotai/Kimi-K2-Instruct",
+                    "tags": "LLM,CHAT,131k",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen3-32B",
+                    "tags": "LLM,CHAT,131k",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen3-14B",
+                    "tags": "LLM,CHAT,131k",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen3-8B",
+                    "tags": "LLM,CHAT,131k",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen3-Reranker-8B",
+                    "tags": "LLM,RE-RANK,33k",
+                    "max_tokens": 33000,
+                    "model_type": "rerank",
+                    "is_tools": false
+                },
+                {
+                    "llm_name": "Qwen/Qwen3-Embedding-8B",
+                    "tags": "LLM,EMBEDDING,33k",
+                    "max_tokens": 33000,
+                    "model_type": "embedding",
+                    "is_tools": false
+                },
+                {
+                    "llm_name": "Qwen/Qwen3-Reranker-4B",
+                    "tags": "LLM,RE-RANK,33k",
+                    "max_tokens": 33000,
+                    "model_type": "rerank",
+                    "is_tools": false
+                },
+                {
+                    "llm_name": "Qwen/Qwen3-Embedding-4B",
+                    "tags": "LLM,EMBEDDING,33k",
+                    "max_tokens": 33000,
+                    "model_type": "embedding",
+                    "is_tools": false
+                },
+                {
+                    "llm_name": "Qwen/Qwen3-Reranker-0.6B",
+                    "tags": "LLM,RE-RANK,33k",
+                    "max_tokens": 33000,
+                    "model_type": "rerank",
+                    "is_tools": false
+                },
+                {
+                    "llm_name": "Qwen/Qwen3-Embedding-0.6B",
+                    "tags": "LLM,EMBEDDING,33k",
+                    "max_tokens": 33000,
+                    "model_type": "embedding",
+                    "is_tools": false
+                },
+                {
+                    "llm_name": "THUDM/GLM-Z1-32B-0414",
+                    "tags": "LLM,CHAT,131k",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "THUDM/GLM-4-32B-0414",
+                    "tags": "LLM,CHAT,33k",
+                    "max_tokens": 33000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "THUDM/GLM-Z1-9B-0414",
+                    "tags": "LLM,CHAT,131k",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "THUDM/GLM-4-9B-0414",
+                    "tags": "LLM,CHAT,33k",
+                    "max_tokens": 33000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen2.5-VL-32B-Instruct",
+                    "tags": "LLM,CHAT,131k",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/QwQ-32B",
+                    "tags": "LLM,CHAT,131k",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen2.5-VL-72B-Instruct",
+                    "tags": "LLM,CHAT,131k",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen2.5-VL-7B-Instruct",
+                    "tags": "LLM,CHAT,33k",
+                    "max_tokens": 33000,
+                    "model_type": "chat",
+                    "is_tools": false
+                },
+                {
+                    "llm_name": "deepseek-ai/DeepSeek-R1-Distill-Qwen-32B",
+                    "tags": "LLM,CHAT,131k",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "deepseek-ai/DeepSeek-R1-Distill-Qwen-14B",
+                    "tags": "LLM,CHAT,131k",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen2.5-Coder-32B-Instruct",
+                    "tags": "LLM,CHAT,33k",
+                    "max_tokens": 33000,
+                    "model_type": "chat",
+                    "is_tools": false
+                },
+                {
+                    "llm_name": "Qwen/Qwen2.5-72B-Instruct-128K",
+                    "tags": "LLM,CHAT,131k",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "deepseek-ai/deepseek-vl2",
+                    "tags": "LLM,CHAT,4k",
+                    "max_tokens": 4000,
+                    "model_type": "chat",
+                    "is_tools": false
+                },
+                {
+                    "llm_name": "Qwen/Qwen2.5-72B-Instruct",
+                    "tags": "LLM,CHAT,33k",
+                    "max_tokens": 33000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen2.5-32B-Instruct",
+                    "tags": "LLM,CHAT,33k",
+                    "max_tokens": 33000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen2.5-14B-Instruct",
+                    "tags": "LLM,CHAT,33k",
+                    "max_tokens": 33000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "Qwen/Qwen2.5-7B-Instruct",
+                    "tags": "LLM,CHAT,33k",
+                    "max_tokens": 33000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "IndexTeam/IndexTTS-2",
+                    "tags": "TTS",
+                    "max_tokens": 1000,
+                    "model_type": "tts",
+                    "is_tools": false
+                }
+            ]
+        },
        {
            "name": "PPIO",
            "logo": "",
--- a/rag/llm/embedding_model.py
+++ b/rag/llm/embedding_model.py
@ -770,14 +770,17 @@ class SILICONFLOWEmbed(Base):
    _FACTORY_NAME = "SILICONFLOW"

    def __init__(self, key, model_name, base_url="https://api.siliconflow.cn/v1/embeddings"):
-        if not base_url:
-            base_url = "https://api.siliconflow.cn/v1/embeddings"
+        normalized_base_url = (base_url or "").strip()
+        if not normalized_base_url:
+            normalized_base_url = "https://api.siliconflow.cn/v1/embeddings"
+        if "/embeddings" not in normalized_base_url:
+            normalized_base_url = urljoin(f"{normalized_base_url.rstrip('/')}/", "embeddings").rstrip("/")
        self.headers = {
            "accept": "application/json",
            "content-type": "application/json",
            "authorization": f"Bearer {key}",
        }
-        self.base_url = base_url
+        self.base_url = normalized_base_url
        self.model_name = model_name

    def encode(self, texts: list):
--- a/rag/llm/rerank_model.py
+++ b/rag/llm/rerank_model.py
@ -274,10 +274,13 @@ class SILICONFLOWRerank(Base):
    _FACTORY_NAME = "SILICONFLOW"

    def __init__(self, key, model_name, base_url="https://api.siliconflow.cn/v1/rerank"):
-        if not base_url:
-            base_url = "https://api.siliconflow.cn/v1/rerank"
+        normalized_base_url = (base_url or "").strip()
+        if not normalized_base_url:
+            normalized_base_url = "https://api.siliconflow.cn/v1/rerank"
+        if "/rerank" not in normalized_base_url:
+            normalized_base_url = urljoin(f"{normalized_base_url.rstrip('/')}/", "rerank").rstrip("/")
        self.model_name = model_name
-        self.base_url = base_url
+        self.base_url = normalized_base_url
        self.headers = {
            "accept": "application/json",
            "content-type": "application/json",
--- a/web/src/hooks/use-llm-request.tsx
+++ b/web/src/hooks/use-llm-request.tsx
@ -271,6 +271,7 @@ export interface IApiKeySavingParams {
  llm_name?: string;
  model_type?: string;
  base_url?: string;
+  source_fid?: string;
  verify?: boolean;
 }

--- a/web/src/locales/bg.ts
+++ b/web/src/locales/bg.ts
@ -1137,6 +1137,8 @@ The above is the content you need to summarize.`,
        'Ако вашият API ключ е от OpenAI, просто го игнорирайте. Всеки друг междинен доставчик ще предостави този base url с API ключа.',
      tongyiBaseUrlTip:
        'За китайски потребители не е необходимо да попълвате или използвайте https://dashscope.aliyuncs.com/compatible-mode/v1. За международни потребители използвайте https://dashscope-intl.aliyuncs.com/compatible-mode/v1',
+      siliconBaseUrlTip:
+        'За китайски потребители не е необходимо да попълвате или използвайте https://api.siliconflow.cn/v1. За международни потребители използвайте https://api.siliconflow.com/v1',
      tongyiBaseUrlPlaceholder:
        '(Само за международни потребители, вижте съвета)',
      minimaxBaseUrlTip:
--- a/web/src/locales/de.ts
+++ b/web/src/locales/de.ts
@ -1177,6 +1177,8 @@ Beispiel: Virtual Hosted Style`,
        'Wenn Ihr API-Schlüssel von OpenAI stammt, ignorieren Sie dies. Andere Zwischenanbieter geben diese Basis-URL mit dem API-Schlüssel an.',
      tongyiBaseUrlTip:
        'Für chinesische Benutzer ist keine Eingabe erforderlich oder verwenden Sie https://dashscope.aliyuncs.com/compatible-mode/v1. Für internationale Benutzer verwenden Sie https://dashscope-intl.aliyuncs.com/compatible-mode/v1',
+      siliconBaseUrlTip:
+        'Für chinesische Benutzer ist keine Eingabe erforderlich oder verwenden Sie https://api.siliconflow.cn/v1. Für internationale Benutzer verwenden Sie https://api.siliconflow.com/v1',
      tongyiBaseUrlPlaceholder:
        '(Nur für internationale Benutzer, bitte Hinweis beachten)',
      minimaxBaseUrlTip:
--- a/web/src/locales/en.ts
+++ b/web/src/locales/en.ts
@ -1172,6 +1172,8 @@ Example: Virtual Hosted Style`,
        'If your API key is from OpenAI, just ignore it. Any other intermediate providers will give this base url with the API key.',
      tongyiBaseUrlTip:
        'For Chinese users, no need to fill in or use https://dashscope.aliyuncs.com/compatible-mode/v1. For international users, use https://dashscope-intl.aliyuncs.com/compatible-mode/v1',
+      siliconBaseUrlTip:
+        'For Chinese users, no need to fill in or use https://api.siliconflow.cn/v1. For international users, use https://api.siliconflow.com/v1',
      tongyiBaseUrlPlaceholder: '(International users only, please see tip)',
      minimaxBaseUrlTip:
        'International users only: use https://api.minimax.io/v1',
--- a/web/src/locales/es.ts
+++ b/web/src/locales/es.ts
@ -358,6 +358,8 @@ export default {
        'Si tu clave API es de OpenAI, ignora esto. Cualquier otro proveedor intermedio proporcionará esta URL base junto con la clave API.',
      tongyiBaseUrlTip:
        'Para usuarios chinos, no es necesario rellenar o usar https://dashscope.aliyuncs.com/compatible-mode/v1. Para usuarios internacionales, usar https://dashscope-intl.aliyuncs.com/compatible-mode/v1',
+      siliconBaseUrlTip:
+        'Para usuarios chinos, no es necesario rellenar o usar https://api.siliconflow.cn/v1. Para usuarios internacionales, usar https://api.siliconflow.com/v1',
      tongyiBaseUrlPlaceholder:
        '(Solo para usuarios internacionales, por favor ver consejo)',
      minimaxBaseUrlTip:
--- a/web/src/locales/fr.ts
+++ b/web/src/locales/fr.ts
@ -542,6 +542,8 @@ export default {
        "Si votre clé API provient d'OpenAI, ignorez ceci. Tout autre fournisseur intermédiaire fournira cette URL de base avec la clé API.",
      tongyiBaseUrlTip:
        'Pour les utilisateurs chinois, pas besoin de remplir ou utiliser https://dashscope.aliyuncs.com/compatible-mode/v1. Pour les utilisateurs internationaux, utilisez https://dashscope-intl.aliyuncs.com/compatible-mode/v1',
+      siliconBaseUrlTip:
+        'Pour les utilisateurs chinois, pas besoin de remplir ou utiliser https://api.siliconflow.cn/v1. Pour les utilisateurs internationaux, utilisez https://api.siliconflow.com/v1',
      tongyiBaseUrlPlaceholder:
        "(Utilisateurs internationaux uniquement, veuillez consulter l'astuce)",
      minimaxBaseUrlTip:
--- a/web/src/locales/id.ts
+++ b/web/src/locales/id.ts
@ -529,6 +529,8 @@ export default {
        'Jika kunci API Anda berasal dari OpenAI, abaikan saja. Penyedia perantara lainnya akan memberikan base url ini dengan kunci API.',
      tongyiBaseUrlTip:
        'Untuk pengguna Tiongkok, tidak perlu diisi atau gunakan https://dashscope.aliyuncs.com/compatible-mode/v1. Untuk pengguna internasional, gunakan https://dashscope-intl.aliyuncs.com/compatible-mode/v1',
+      siliconBaseUrlTip:
+        'Untuk pengguna Tiongkok, tidak perlu diisi atau gunakan https://api.siliconflow.cn/v1. Untuk pengguna internasional, gunakan https://api.siliconflow.com/v1',
      tongyiBaseUrlPlaceholder:
        '(Hanya untuk pengguna internasional, silakan lihat tip)',
      minimaxBaseUrlTip:
--- a/web/src/locales/it.ts
+++ b/web/src/locales/it.ts
@ -770,6 +770,8 @@ Quanto sopra è il contenuto che devi riassumere.`,
      baseUrl: 'URL Base',
      baseUrlTip:
        'Se la tua chiave API è da OpenAI, ignoralo. Qualsiasi altro fornitore intermedio fornirà questo URL base con la chiave API.',
+      siliconBaseUrlTip:
+        'For Chinese users, no need to fill in or use https://api.siliconflow.cn/v1. For international users, use https://api.siliconflow.com/v1',
      modify: 'Modifica',
      systemModelSettings: 'Imposta modelli predefiniti',
      chatModel: 'LLM',
--- a/web/src/locales/ja.ts
+++ b/web/src/locales/ja.ts
@ -569,6 +569,8 @@ export default {
        'APIキーがOpenAIからのものであれば無視してください。他の中間プロバイダーはAPIキーと共にこのベースURLを提供します。',
      tongyiBaseUrlTip:
        '中国ユーザーの場合、記入不要または https://dashscope.aliyuncs.com/compatible-mode/v1 を使用してください。国際ユーザーは https://dashscope-intl.aliyuncs.com/compatible-mode/v1 を使用してください',
+      siliconBaseUrlTip:
+        '中国ユーザーの場合、入力不要または https://api.siliconflow.cn/v1 を使用してください。国際ユーザーは https://api.siliconflow.com/v1 を使用してください',
      tongyiBaseUrlPlaceholder: '（国際ユーザーのみ、ヒントをご覧ください）',
      minimaxBaseUrlTip:
        '国際ユーザーのみ：https://api.minimax.io/v1 を使用してください。',
--- a/web/src/locales/pt-br.ts
+++ b/web/src/locales/pt-br.ts
@ -524,6 +524,8 @@ export default {
        'Se sua chave da API for do OpenAI, ignore isso. Outros provedores intermediários fornecerão essa URL base com a chave da API.',
      tongyiBaseUrlTip:
        'Para usuários chineses, não é necessário preencher ou usar https://dashscope.aliyuncs.com/compatible-mode/v1. Para usuários internacionais, use https://dashscope-intl.aliyuncs.com/compatible-mode/v1',
+      siliconBaseUrlTip:
+        'Para usuários chineses, não é necessário preencher ou usar https://api.siliconflow.cn/v1. Para usuários internacionais, use https://api.siliconflow.com/v1',
      tongyiBaseUrlPlaceholder:
        '(Apenas para usuários internacionais, consulte a dica)',
      minimaxBaseUrlTip:
--- a/web/src/locales/ru.ts
+++ b/web/src/locales/ru.ts
@ -894,6 +894,8 @@ export default {
        'Если ваш API ключ от OpenAI, просто проигнорируйте это. Любые другие промежуточные провайдеры дадут этот базовый url вместе с API ключом.',
      tongyiBaseUrlTip:
        'Для китайских пользователей не нужно заполнять или используйте https://dashscope.aliyuncs.com/compatible-mode/v1. Для международных пользователей используйте https://dashscope-intl.aliyuncs.com/compatible-mode/v1',
+      siliconBaseUrlTip:
+        'Для китайских пользователей не нужно заполнять или используйте https://api.siliconflow.cn/v1. Для международных пользователей используйте https://api.siliconflow.com/v1',
      tongyiBaseUrlPlaceholder:
        '(Только для международных пользователей, см. подсказку)',
      minimaxBaseUrlTip:
--- a/web/src/locales/vi.ts
+++ b/web/src/locales/vi.ts
@ -573,6 +573,8 @@ export default {
      baseUrl: 'Base-Url',
      baseUrlTip:
        'Nếu khóa API của bạn từ OpenAI, chỉ cần bỏ qua nó. Bất kỳ nhà cung cấp trung gian nào khác sẽ cung cấp URL cơ sở này với khóa API.',
+      siliconBaseUrlTip:
+        'For Chinese users, no need to fill in or use https://api.siliconflow.cn/v1. For international users, use https://api.siliconflow.com/v1',
      minimaxBaseUrlTip:
        'Chỉ người dùng quốc tế: dùng https://api.minimax.io/v1.',
      minimaxBaseUrlPlaceholder:
--- a/web/src/locales/zh-traditional.ts
+++ b/web/src/locales/zh-traditional.ts
@ -626,6 +626,8 @@ export default {
        '如果您的 API 密鑰來自 OpenAI，請忽略它。任何其他中間提供商都會提供帶有 API 密鑰的基本 URL。',
      tongyiBaseUrlTip:
        '中國用戶無需填寫或使用 https://dashscope.aliyuncs.com/compatible-mode/v1。國際用戶請使用 https://dashscope-intl.aliyuncs.com/compatible-mode/v1',
+      siliconBaseUrlTip:
+        '中國用戶無需填寫或使用 https://api.siliconflow.cn/v1。國際用戶請使用 https://api.siliconflow.com/v1',
      tongyiBaseUrlPlaceholder: '（僅國際用戶，請參閱提示）',
      minimaxBaseUrlTip: '僅國際用戶：使用 https://api.minimax.io/v1。',
      minimaxBaseUrlPlaceholder: '（僅國際用戶填寫 https://api.minimax.io/v1）',
--- a/web/src/locales/zh.ts
+++ b/web/src/locales/zh.ts
@ -991,6 +991,8 @@ General：实体和关系提取提示来自 GitHub - microsoft/graphrag：基于
        '如果您的 API 密钥来自 OpenAI，请忽略它。 任何其他中间提供商都会提供带有 API 密钥的基本 URL。',
      tongyiBaseUrlTip:
        '对于中国用户，不需要填写或使用 https://dashscope.aliyuncs.com/compatible-mode/v1。对于国际用户，使用 https://dashscope-intl.aliyuncs.com/compatible-mode/v1。',
+      siliconBaseUrlTip:
+        '对于中国用户，不需要填写或使用 https://api.siliconflow.cn/v1。对于国际用户，使用 https://api.siliconflow.com/v1。',
      tongyiBaseUrlPlaceholder: '(仅国际用户需要)',
      minimaxBaseUrlTip: '仅国际用户：使用 https://api.minimax.io/v1。',
      minimaxBaseUrlPlaceholder: '(仅国际用户填写 https://api.minimax.io/v1)',
--- a/web/src/pages/user-setting/setting-model/hooks.tsx
+++ b/web/src/pages/user-setting/setting-model/hooks.tsx
@ -43,11 +43,20 @@ export const useSubmitApiKey = () => {
      if (!isVerify) {
        setSaveLoading(true);
      }
-      const ret = await saveApiKey({
+      const payload: IApiKeySavingParams = {
        ...savingParams,
        ...postBody,
        verify: isVerify,
-      });
+      };
+      if (savingParams.llm_factory === LLMFactory.SILICONFLOW) {
+        payload.source_fid = (postBody.base_url || '')
+          .toLowerCase()
+          .includes('api.siliconflow.com')
+          ? 'siliconflow_intl'
+          : LLMFactory.SILICONFLOW;
+      }
+
+      const ret = await saveApiKey(payload);
      if (!isVerify) {
        setSaveLoading(false);
        if (ret.code === 0) {
--- a/web/src/pages/user-setting/setting-model/modal/api-key-modal/index.tsx
+++ b/web/src/pages/user-setting/setting-model/modal/api-key-modal/index.tsx
@ -41,6 +41,7 @@ const modelsWithBaseUrl = [
  LLMFactory.AzureOpenAI,
  LLMFactory.TongYiQianWen,
  LLMFactory.MiniMax,
+  LLMFactory.SILICONFLOW,
 ];

 const ApiKeyModal = ({
@ -127,7 +128,9 @@ const ApiKeyModal = ({
                        ? t('minimaxBaseUrlTip')
                        : llmFactory === LLMFactory.TongYiQianWen
                          ? t('tongyiBaseUrlTip')
-                          : t('baseUrlTip')
+                          : llmFactory === LLMFactory.SILICONFLOW
+                            ? t('siliconBaseUrlTip')
+                            : t('baseUrlTip')
                    }
                  >
                    {t('baseUrl')}
@ -140,7 +143,9 @@ const ApiKeyModal = ({
                          ? t('tongyiBaseUrlPlaceholder')
                          : llmFactory === LLMFactory.MiniMax
                            ? t('minimaxBaseUrlPlaceholder')
-                            : 'https://api.openai.com/v1'
+                            : llmFactory === LLMFactory.SILICONFLOW
+                              ? 'https://api.siliconflow.cn/v1'
+                              : 'https://api.openai.com/v1'
                      }
                      onKeyDown={handleKeyDown}
                      className="w-full"