Feat/support 302ai provider (#8742)

### What problem does this PR solve?

Support 302.AI provider.

### Type of change

- [x] New Feature (non-breaking change which adds functionality)

conf/llm_factories.json

@@ -160,7 +160,6 @@
                     "max_tokens": 131072,
                     "model_type": "chat",
                     "is_tools": true
-
                 },
                 {
                     "llm_name": "grok-3-fast",
@@ -168,7 +167,6 @@
                     "max_tokens": 131072,
                     "model_type": "chat",
                     "is_tools": true
-
                 },
                 {
                     "llm_name": "grok-3-mini",
@@ -176,7 +174,6 @@
                     "max_tokens": 131072,
                     "model_type": "chat",
                     "is_tools": true
-
                 },
                 {
                     "llm_name": "grok-3-mini-mini-fast",
@@ -184,7 +181,6 @@
                     "max_tokens": 131072,
                     "model_type": "chat",
                     "is_tools": true
-
                 },
                 {
                     "llm_name": "grok-2-vision",
@@ -192,7 +188,6 @@
                     "max_tokens": 32768,
                     "model_type": "image2text",
                     "is_tools": true
-
                 }
             ]
         },
@@ -851,7 +846,7 @@
         {
             "name": "BAAI",
             "logo": "",
-           "tags": "TEXT EMBEDDING",
+            "tags": "TEXT EMBEDDING",
             "status": "1",
             "llm": [
                 {
@@ -4130,6 +4125,175 @@
                     "model_type": "embedding"
                 }
             ]
+        },
+        {
+            "name": "302.AI",
+            "logo": "",
+            "tags": "LLM,TEXT EMBEDDING,TEXT RE-RANK,IMAGE2TEXT",
+            "status": "1",
+            "llm": [
+                {
+                    "llm_name": "deepseek-chat",
+                    "tags": "LLM,CHAT",
+                    "max_tokens": 32000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "gpt-4o",
+                    "tags": "LLM,CHAT,128k",
+                    "max_tokens": 128000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "chatgpt-4o-latest",
+                    "tags": "LLM,CHAT,128k",
+                    "max_tokens": 128000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "llama3.3-70b",
+                    "tags": "LLM,CHAT,128k",
+                    "max_tokens": 128000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "deepseek-reasoner",
+                    "tags": "LLM,CHAT,64k",
+                    "max_tokens": 64000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "gemini-2.0-flash",
+                    "tags": "LLM,CHAT",
+                    "max_tokens": 1000000,
+                    "model_type": "image2text",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "claude-3-7-sonnet-20250219",
+                    "tags": "LLM,CHAT",
+                    "max_tokens": 200000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "claude-3-7-sonnet-latest",
+                    "tags": "LLM,CHAT",
+                    "max_tokens": 200000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "grok-3-beta",
+                    "tags": "LLM,CHAT",
+                    "max_tokens": 131072,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "grok-3-mini-beta",
+                    "tags": "LLM,CHAT",
+                    "max_tokens": 131072,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "gpt-4.1",
+                    "tags": "LLM,CHAT",
+                    "max_tokens": 1000000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "o3",
+                    "tags": "LLM,CHAT",
+                    "max_tokens": 200000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "o4-mini",
+                    "tags": "LLM,CHAT",
+                    "max_tokens": 200000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "qwen3-235b-a22b",
+                    "tags": "LLM,CHAT,128k",
+                    "max_tokens": 128000,
+                    "model_type": "chat",
+                    "is_tools": false
+                },
+                {
+                    "llm_name": "qwen3-32b",
+                    "tags": "LLM,CHAT,128k",
+                    "max_tokens": 128000,
+                    "model_type": "chat",
+                    "is_tools": false
+                },
+                {
+                    "llm_name": "gemini-2.5-pro-preview-05-06",
+                    "tags": "LLM,CHAT",
+                    "max_tokens": 1000000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "llama-4-maverick",
+                    "tags": "LLM,CHAT,128k",
+                    "max_tokens": 128000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "gemini-2.5-flash",
+                    "tags": "LLM,CHAT",
+                    "max_tokens": 1000000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "claude-sonnet-4-20250514",
+                    "tags": "LLM,CHAT",
+                    "max_tokens": 200000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "claude-opus-4-20250514",
+                    "tags": "LLM,CHAT",
+                    "max_tokens": 200000,
+                    "model_type": "image2text",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "gemini-2.5-pro",
+                    "tags": "LLM,CHAT",
+                    "max_tokens": 1000000,
+                    "model_type": "image2text",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "jina-clip-v2",
+                    "tags": "TEXT EMBEDDING,TEXT RE-RANK",
+                    "max_tokens": 8192,
+                    "model_type": "embedding",
+                    "is_tools": false
+                },
+                {
+                    "llm_name": "jina-reranker-m0",
+                    "tags": "TEXT EMBEDDING,TEXT RE-RANK",
+                    "max_tokens": 10240,
+                    "model_type": "rerank",
+                    "is_tools": false
+                }
+            ]
         }
     ]
 }

docs/references/supported_models.mdx

@@ -4,6 +4,7 @@ slug: /supported_models
 ---
 
 # Supported models
+
 import APITable from '@site/src/components/APITable';
 
 A complete list of models supported by RAGFlow, which will continue to expand.
@@ -63,6 +64,7 @@ A complete list of models supported by RAGFlow, which will continue to expand.
 | ZHIPU-AI              | :heavy_check_mark: | :heavy_check_mark: |                    | :heavy_check_mark: |                    |                    |
 | 01.AI                 | :heavy_check_mark: |                    |                    |                    |                    |                    |
 | DeepInfra             | :heavy_check_mark: | :heavy_check_mark: |                    |                    | :heavy_check_mark: | :heavy_check_mark: |
+| 302.AI                | :heavy_check_mark: | :heavy_check_mark: | :heavy_check_mark: | :heavy_check_mark: |                    |                    |
 
 ```mdx-code-block
 </APITable>

rag/llm/chat_model.py

@@ -1724,8 +1724,6 @@ class GPUStackChat(Base):
             raise ValueError("Local llm url cannot be None")
         base_url = urljoin(base_url, "v1")
         super().__init__(key, model_name, base_url, **kwargs)
-
-
 class DeepInfraChat(Base):
     _FACTORY_NAME = "DeepInfra"
 
@@ -1733,3 +1731,12 @@ class DeepInfraChat(Base):
         if not base_url:
             base_url = "https://api.deepinfra.com/v1/openai"
         super().__init__(key, model_name, base_url, **kwargs)
+        
+
+class Ai302Chat(Base):
+    _FACTORY_NAME = "302.AI"
+
+    def __init__(self, key, model_name, base_url="https://api.302.ai/v1", **kwargs):
+        if not base_url:
+            base_url = "https://api.302.ai/v1"
+        super().__init__(key, model_name, base_url, **kwargs)

rag/llm/embedding_model.py

@@ -931,7 +931,6 @@ class GiteeEmbed(SILICONFLOWEmbed):
             base_url = "https://ai.gitee.com/v1/embeddings"
         super().__init__(key, model_name, base_url)
         
-
 class DeepInfraEmbed(OpenAIEmbed):
     _FACTORY_NAME = "DeepInfra"
 
@@ -939,3 +938,12 @@ class DeepInfraEmbed(OpenAIEmbed):
         if not base_url:
             base_url = "https://api.deepinfra.com/v1/openai"
         super().__init__(key, model_name, base_url)
+
+
+class Ai302Embed(Base):
+    _FACTORY_NAME = "302.AI"
+
+    def __init__(self, key, model_name, base_url="https://api.302.ai/v1/embeddings"):
+        if not base_url:
+            base_url = "https://api.302.ai/v1/embeddings"
+        super().__init__(key, model_name, base_url)

rag/llm/rerank_model.py

@@ -617,3 +617,12 @@ class GiteeRerank(JinaRerank):
         if not base_url:
             base_url = "https://ai.gitee.com/v1/rerank"
         super().__init__(key, model_name, base_url)
+
+
+class Ai302Rerank(Base):
+    _FACTORY_NAME = "302.AI"
+
+    def __init__(self, key, model_name, base_url="https://api.302.ai/v1/rerank"):
+        if not base_url:
+            base_url = "https://api.302.ai/v1/rerank"
+        super().__init__(key, model_name, base_url)

rag/llm/sequence2txt_model.py

@@ -208,7 +208,6 @@ class GiteeSeq2txt(Base):
         self.client = OpenAI(api_key=key, base_url=base_url)
         self.model_name = model_name
 
-
 class DeepInfraSeq2txt(Base):
     _FACTORY_NAME = "DeepInfra"
 

rag/llm/tts_model.py

@@ -383,7 +383,6 @@ class SILICONFLOWTTS(Base):
             if chunk:
                 yield chunk
 
-
 class DeepInfraTTS(OpenAITTS):
     _FACTORY_NAME = "DeepInfra"
 

web/src/assets/svg/llm/ai302.svg

@@ -0,0 +1 @@
+<svg height="1em" style="flex:none;line-height:1" viewBox="0 0 24 24" width="1em" xmlns="http://www.w3.org/2000/svg"><title>302.AI</title><path d="M13.086 23.25c5.614 0 10.164-4.559 10.164-10.182 0-5.624-4.55-10.182-10.164-10.182-5.613 0-10.163 4.558-10.163 10.182 0 5.623 4.55 10.182 10.163 10.182z" fill="#3F3FAA"></path><path d="M10.914 21.114c5.613 0 10.163-4.559 10.163-10.182S16.527.75 10.914.75C5.3.75.75 5.309.75 10.932S5.3 21.114 10.914 21.114z" fill="#8E47FF"></path><path d="M10.755 17.708c-.722 0-1.416-.24-1.995-.69a3.25 3.25 0 01-1.23-2.177 3.248 3.248 0 01-.006-.782c-.201.04-.407.056-.618.056a3.265 3.265 0 01-3.261-3.262A3.27 3.27 0 017.65 7.67a3.27 3.27 0 013.241-3.638 3.266 3.266 0 013.242 3.577 3.269 3.269 0 012.694 5.693 3.227 3.227 0 01-2.365.782 3.26 3.26 0 01-.466-.066c0 .008.005.014.005.02a3.254 3.254 0 01-.664 2.41 3.235 3.235 0 01-2.583 1.259zm-1.808-4.313c-.228.397-.32.847-.263 1.304a2.092 2.092 0 002.335 1.826 2.086 2.086 0 001.398-.791 2.08 2.08 0 00.425-1.548 2.091 2.091 0 00-.405-1.004 3.253 3.253 0 01-.39-.462.58.58 0 11.947-.675c.044.062.088.117.137.173a.61.61 0 01.111.101l.056.071a2.096 2.096 0 003.49-1.405 2.096 2.096 0 00-1.93-2.248 2.076 2.076 0 00-1.251.304.579.579 0 01-.648.061.59.59 0 01-.233-.796A2.102 2.102 0 0010.888 5.2 2.1 2.1 0 009.14 8.457l.03.056c.61.594.993 1.42.993 2.339A3.273 3.273 0 018.947 13.4v-.005zM6.901 8.752a2.098 2.098 0 00-2.092 2.1c0 1.16.936 2.101 2.092 2.101a2.1 2.1 0 000-4.201z" fill="#fff"></path></svg>

web/src/constants/llm.ts

@@ -50,6 +50,7 @@ export enum LLMFactory {
   GPUStack = 'GPUStack',
   VLLM = 'VLLM',
   GiteeAI = 'GiteeAI',
+  Ai302 = '302.AI',
   DeepInfra = 'DeepInfra',
 }
 
@@ -106,5 +107,6 @@ export const IconMap = {
   [LLMFactory.GPUStack]: 'gpustack',
   [LLMFactory.VLLM]: 'vllm',
   [LLMFactory.GiteeAI]: 'gitee-ai',
+  [LLMFactory.Ai302]: 'ai302',
   [LLMFactory.DeepInfra]: 'deepinfra',
 };

web/src/utils/common-util.ts

@@ -43,6 +43,7 @@ const orderFactoryList = [
   LLMFactory.ZhipuAI,
   LLMFactory.Ollama,
   LLMFactory.Xinference,
+  LLMFactory.Ai302,
 ];
 
 export const sortLLmFactoryListBySpecifiedOrder = (list: IFactory[]) => {