@lobehub/chat 1.19.5 → 1.19.7

package/CHANGELOG.md

@@ -2,6 +2,56 @@
 
 # Changelog
 
+### [Version 1.19.7](https://github.com/lobehub/lobe-chat/compare/v1.19.6...v1.19.7)
+
+<sup>Released on **2024-09-19**</sup>
+
+#### 💄 Styles
+
+- **misc**: Add siliconflow qwen2.5 model.
+
+<br/>
+
+<details>
+<summary><kbd>Improvements and Fixes</kbd></summary>
+
+#### Styles
+
+- **misc**: Add siliconflow qwen2.5 model, closes [#4024](https://github.com/lobehub/lobe-chat/issues/4024) ([06ffd99](https://github.com/lobehub/lobe-chat/commit/06ffd99))
+
+</details>
+
+<div align="right">
+
+[![](https://img.shields.io/badge/-BACK_TO_TOP-151515?style=flat-square)](#readme-top)
+
+</div>
+
+### [Version 1.19.6](https://github.com/lobehub/lobe-chat/compare/v1.19.5...v1.19.6)
+
+<sup>Released on **2024-09-19**</sup>
+
+#### ♻ Code Refactoring
+
+- **misc**: Refactor the tts route url.
+
+<br/>
+
+<details>
+<summary><kbd>Improvements and Fixes</kbd></summary>
+
+#### Code refactoring
+
+- **misc**: Refactor the tts route url, closes [#4030](https://github.com/lobehub/lobe-chat/issues/4030) ([60dcf19](https://github.com/lobehub/lobe-chat/commit/60dcf19))
+
+</details>
+
+<div align="right">
+
+[![](https://img.shields.io/badge/-BACK_TO_TOP-151515?style=flat-square)](#readme-top)
+
+</div>
+
 ### [Version 1.19.5](https://github.com/lobehub/lobe-chat/compare/v1.19.4...v1.19.5)
 
 <sup>Released on **2024-09-19**</sup>

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@lobehub/chat",
-  "version": "1.19.5",
+  "version": "1.19.7",
   "description": "Lobe Chat - an open-source, high-performance chatbot framework that supports speech synthesis, multimodal, and extensible Function Call plugin system. Supports one-click free deployment of your private ChatGPT/LLM web application.",
   "keywords": [
     "framework",

package/src/app/api/openai/createBizOpenAI/index.ts

@@ -8,6 +8,7 @@ import { checkAuth } from './auth';
 import { createOpenai } from './createOpenai';
 
 /**
+ * @deprecated
  * createOpenAI Instance with Auth and azure openai support
  * if auth not pass ,just return error response
  */

package/src/app/{api/openai/tts → webapi/tts/openai}/route.ts

@@ -28,6 +28,7 @@ export const preferredRegion = [
 export const POST = async (req: Request) => {
   const payload = (await req.json()) as OpenAITTSPayload;
 
+  // need to be refactored with jwt auth mode
   const openaiOrErrResponse = createBizOpenAI(req);
 
   // if resOrOpenAI is a Response, it means there is an error,just return it

package/src/config/modelProviders/qwen.ts

@@ -43,9 +43,33 @@ const Qwen: ModelProviderCard = {
       },
       tokens: 32_768,
     },
+    {
+      description:
+        '通义千问数学模型是专门用于数学解题的语言模型。',
+      displayName: 'Qwen Math Turbo',
+      id: 'qwen-math-turbo-latest',
+      pricing: {
+        currency: 'CNY',
+        input: 2,
+        output: 6,
+      },
+      tokens: 4096,
+    },
+    {
+      description:
+        '通义千问数学模型是专门用于数学解题的语言模型。',
+      displayName: 'Qwen Math Plus',
+      id: 'qwen-math-plus-latest',
+      pricing: {
+        currency: 'CNY',
+        input: 4,
+        output: 12,
+      },
+      tokens: 4096,
+    },
     {
       description: '通义千问代码模型。',
-      displayName: 'Qwen Coder',
+      displayName: 'Qwen Coder Turbo',
       id: 'qwen-coder-turbo-latest',
       pricing: {
         currency: 'CNY',
@@ -94,35 +118,16 @@ const Qwen: ModelProviderCard = {
       tokens: 32_000,
       vision: true,
     },
-    {
-      description:
-        '通义千问数学模型是专门用于数学解题的语言模型。',
-      displayName: 'Qwen Math Turbo',
-      id: 'qwen-math-turbo-latest',
-      pricing: {
-        currency: 'CNY',
-        input: 2,
-        output: 6,
-      },
-      tokens: 4096,
-    },
-    {
-      description:
-        '通义千问数学模型是专门用于数学解题的语言模型。',
-      displayName: 'Qwen Math Plus',
-      id: 'qwen-math-plus-latest',
-      pricing: {
-        currency: 'CNY',
-        input: 4,
-        output: 12,
-      },
-      tokens: 4096,
-    },
     {
       description: '通义千问2.5对外开源的7B规模的模型。',
       displayName: 'Qwen2.5 7B',
       functionCall: true,
       id: 'qwen2.5-7b-instruct',
+      pricing: {
+        currency: 'CNY',
+        input: 1,
+        output: 2,
+      },
       tokens: 131_072,
     },
     {
@@ -130,6 +135,11 @@ const Qwen: ModelProviderCard = {
       displayName: 'Qwen2.5 14B',
       functionCall: true,
       id: 'qwen2.5-14b-instruct',
+      pricing: {
+        currency: 'CNY',
+        input: 2,
+        output: 6,
+      },
       tokens: 131_072,
     },
     {
@@ -137,6 +147,11 @@ const Qwen: ModelProviderCard = {
       displayName: 'Qwen2.5 32B',
       functionCall: true,
       id: 'qwen2.5-32b-instruct',
+      pricing: {
+        currency: 'CNY',
+        input: 3.5,
+        output: 7,
+      },
       tokens: 131_072,
     },
     {
@@ -144,15 +159,13 @@ const Qwen: ModelProviderCard = {
       displayName: 'Qwen2.5 72B',
       functionCall: true,
       id: 'qwen2.5-72b-instruct',
+      pricing: {
+        currency: 'CNY',
+        input: 4,
+        output: 12,
+      },
       tokens: 131_072,
     },
-    {
-      description: '通义千问2对外开源的57B规模14B激活参数的MOE模型。',
-      displayName: 'Qwen2 57B A14B MoE',
-      functionCall: true,
-      id: 'qwen2-57b-a14b-instruct',
-      tokens: 65_536,
-    },
     {
       description: 'Qwen-Math 模型具有强大的数学解题能力。',
       displayName: 'Qwen2.5 Math 1.5B',
@@ -168,12 +181,22 @@ const Qwen: ModelProviderCard = {
       description: 'Qwen-Math 模型具有强大的数学解题能力。',
       displayName: 'Qwen2.5 Math 7B',
       id: 'qwen2.5-math-7b-instruct',
+      pricing: {
+        currency: 'CNY',
+        input: 1,
+        output: 2,
+      },
       tokens: 4096,
     },
     {
       description: 'Qwen-Math 模型具有强大的数学解题能力。',
       displayName: 'Qwen2.5 Math 72B',
       id: 'qwen2.5-math-72b-instruct',
+      pricing: {
+        currency: 'CNY',
+        input: 4,
+        output: 12,
+      },
       tokens: 4096,
     },
     {
@@ -191,12 +214,22 @@ const Qwen: ModelProviderCard = {
       description: '通义千问代码模型开源版。',
       displayName: 'Qwen2.5 Coder 7B',
       id: 'qwen2.5-coder-7b-instruct',
+      pricing: {
+        currency: 'CNY',
+        input: 1,
+        output: 2,
+      },
       tokens: 131_072,
     },
     {
       description: '以 Qwen-7B 语言模型初始化，添加图像模型，图像输入分辨率为448的预训练模型。',
       displayName: 'Qwen VL',
       id: 'qwen-vl-v1',
+      pricing: {
+        currency: 'CNY',
+        input: 0,
+        output: 0,
+      },
       tokens: 8000,
       vision: true,
     },
@@ -204,11 +237,16 @@ const Qwen: ModelProviderCard = {
       description: '通义千问VL支持灵活的交互方式，包括多图、多轮问答、创作等能力的模型。',
       displayName: 'Qwen VL Chat',
       id: 'qwen-vl-chat-v1',
+      pricing: {
+        currency: 'CNY',
+        input: 0,
+        output: 0,
+      },
       tokens: 8000,
       vision: true,
     },
   ],
-  checkModel: 'qwen-turbo',
+  checkModel: 'qwen-turbo-latest',
   description:
     '通义千问是阿里云自主研发的超大规模语言模型，具有强大的自然语言理解和生成能力。它可以回答各种问题、创作文字内容、表达观点看法、撰写代码等，在多个领域发挥作用。',
   disableBrowserRequest: true,

package/src/config/modelProviders/siliconcloud.ts

@@ -4,132 +4,160 @@ import { ModelProviderCard } from '@/types/llm';
 const SiliconCloud: ModelProviderCard = {
   chatModels: [
     {
-      description: 'Qwen2 是全新的大型语言模型系列，旨在优化指令式任务的处理。',
-      displayName: 'Qwen2 1.5B',
-      id: 'Qwen/Qwen2-1.5B-Instruct',
+      description: 'DeepSeek V2.5 集合了先前版本的优秀特征，增强了通用和编码能力。',
+      displayName: 'DeepSeek V2.5',
+      enabled: true,
+      id: 'deepseek-ai/DeepSeek-V2.5',
+      pricing: {
+        currency: 'CNY',
+        input: 1.33,
+        output: 1.33,
+      },
       tokens: 32_768,
     },
     {
-      description: 'Qwen2 是全新的大型语言模型系列，具有更强的理解和生成能力。',
-      displayName: 'Qwen2 7B',
-      enabled: true,
-      id: 'Qwen/Qwen2-7B-Instruct',
+      description: 'DeepSeek V2 具备67亿参数，支持英中文本处理。',
+      displayName: 'DeepSeek V2 Chat',
+      id: 'deepseek-ai/DeepSeek-V2-Chat',
+      pricing: {
+        currency: 'CNY',
+        input: 1.33,
+        output: 1.33,
+      },
       tokens: 32_768,
     },
     {
-      description: 'Qwen2 是全新的系列，57B A14B 型号在指令任务中表现卓越。',
-      displayName: 'Qwen2 57B A14B',
-      id: 'Qwen/Qwen2-57B-A14B-Instruct',
+      description: 'DeepSeek Coder V2 为代码任务设计, 专注于高效的代码生成。',
+      displayName: 'DeepSeek V2 Coder',
+      id: 'deepseek-ai/DeepSeek-Coder-V2-Instruct',
+      pricing: {
+        currency: 'CNY',
+        input: 1.33,
+        output: 1.33,
+      },
       tokens: 32_768,
     },
     {
-      description: 'Qwen2 是先进的通用语言模型，支持多种指令类型。',
-      displayName: 'Qwen2 72B',
-      enabled: true,
-      id: 'Qwen/Qwen2-72B-Instruct',
+      description: 'DeepSeek 67B 是为高复杂性对话训练的先进模型。',
+      displayName: 'DeepSeek LLM 67B',
+      id: 'deepseek-ai/deepseek-llm-67b-chat',
+      pricing: {
+        currency: 'CNY',
+        input: 1,
+        output: 1,
+      },
       tokens: 32_768,
     },
     {
-      description: 'Qwen2-Math 专注于数学领域的问题求解，为高难度题提供专业解答。',
-      displayName: 'Qwen2 Math 72B',
+      description: 'Qwen2.5 是全新的大型语言模型系列，旨在优化指令式任务的处理。',
+      displayName: 'Qwen2.5 7B',
       enabled: true,
-      id: 'Qwen/Qwen2-Math-72B-Instruct',
+      id: 'Qwen/Qwen2.5-7B-Instruct',
+      pricing: {
+        currency: 'CNY',
+        input: 0,
+        output: 0,
+      },
       tokens: 32_768,
     },
     {
-      description: 'Qwen1.5 通过结合高级预训练和微调提升对话表达能力。',
-      displayName: 'Qwen1.5 7B',
-      id: 'Qwen/Qwen1.5-7B-Chat',
+      description: 'Qwen2.5 是全新的大型语言模型系列，旨在优化指令式任务的处理。',
+      displayName: 'Qwen2.5 14B',
+      id: 'Qwen/Qwen2.5-14B-Instruct',
+      pricing: {
+        currency: 'CNY',
+        input: 0.7,
+        output: 0.7,
+      },
       tokens: 32_768,
     },
     {
-      description: 'Qwen1.5 通过大规模数据集训练，擅长复杂的语言任务。',
-      displayName: 'Qwen1.5 14B',
-      id: 'Qwen/Qwen1.5-14B-Chat',
+      description: 'Qwen2.5 是全新的大型语言模型系列，旨在优化指令式任务的处理。',
+      displayName: 'Qwen2.5 32B',
+      id: 'Qwen/Qwen2.5-32B-Instruct',
+      pricing: {
+        currency: 'CNY',
+        input: 1.26,
+        output: 1.26,
+      },
       tokens: 32_768,
     },
     {
-      description: 'Qwen1.5 具备多领域问答和文本生成的能力。',
-      displayName: 'Qwen1.5 32B',
-      id: 'Qwen/Qwen1.5-32B-Chat',
+      description: 'Qwen2.5 是全新的大型语言模型系列，具有更强的理解和生成能力。',
+      displayName: 'Qwen2.5 72B',
+      enabled: true,
+      id: 'Qwen/Qwen2.5-72B-Instruct',
+      pricing: {
+        currency: 'CNY',
+        input: 4.13,
+        output: 4.13,
+      },
       tokens: 32_768,
     },
     {
-      description: '作为Qwen2 的测试版，Qwen1.5 使用大规模数据实现了更精确的对话功能。',
-      displayName: 'Qwen1.5 110B',
-      id: 'Qwen/Qwen1.5-110B-Chat',
+      description: 'Qwen2-Math 专注于数学领域的问题求解，为高难度题提供专业解答。',
+      displayName: 'Qwen2 Math 72B',
+      id: 'Qwen/Qwen2-Math-72B-Instruct',
+      pricing: {
+        currency: 'CNY',
+        input: 4.13,
+        output: 4.13,
+      },
       tokens: 32_768,
     },
-    {
-      description: 'Yi-1.5 是Yi系列的进化版本，拥有高质量的预训练和丰富的微调数据。',
-      displayName: 'Yi-1.5 6B',
-      id: '01-ai/Yi-1.5-6B-Chat',
-      tokens: 4096,
-    },
     {
       description: 'Yi-1.5 9B 支持16K Tokens, 提供高效、流畅的语言生成能力。',
       displayName: 'Yi-1.5 9B',
-      enabled: true,
       id: '01-ai/Yi-1.5-9B-Chat-16K',
+      pricing: {
+        currency: 'CNY',
+        input: 0,
+        output: 0,
+      },
       tokens: 16_384,
     },
     {
       description: 'Yi-1.5 34B, 以丰富的训练样本在行业应用中提供优越表现。',
       displayName: 'Yi-1.5 34B',
-      enabled: true,
       id: '01-ai/Yi-1.5-34B-Chat-16K',
+      pricing: {
+        currency: 'CNY',
+        input: 1.26,
+        output: 1.26,
+      },
       tokens: 16_384,
     },
-    {
-      description: '作为双语会话语言模型, ChatGLM3能处理中英文转换任务。',
-      displayName: 'GLM-3 6B',
-      id: 'THUDM/chatglm3-6b',
-      tokens: 32_768,
-    },
     {
       description: 'GLM-4 9B 开放源码版本，为会话应用提供优化后的对话体验。',
       displayName: 'GLM-4 9B',
-      enabled: true,
       id: 'THUDM/glm-4-9b-chat',
+      pricing: {
+        currency: 'CNY',
+        input: 0,
+        output: 0,
+      },
       tokens: 32_768,
     },
     {
       description: 'InternLM2.5 提供多场景下的智能对话解决方案。',
       displayName: 'Internlm 2.5 7B',
-      enabled: true,
       id: 'internlm/internlm2_5-7b-chat',
+      pricing: {
+        currency: 'CNY',
+        input: 0,
+        output: 0,
+      },
       tokens: 32_768,
     },
     {
       description: '创新的开源模型InternLM2.5，通过大规模的参数提高了对话智能。',
       displayName: 'Internlm 2.5 20B',
-      enabled: true,
       id: 'internlm/internlm2_5-20b-chat',
-      tokens: 32_768,
-    },
-    {
-      description: 'DeepSeek V2.5 集合了先前版本的优秀特征，增强了通用和编码能力。',
-      displayName: 'DeepSeek V2.5',
-      enabled: true,
-      id: 'deepseek-ai/DeepSeek-V2.5',
-      tokens: 32_768,
-    },
-    {
-      description: 'DeepSeek V2 具备67亿参数，支持英中文本处理。',
-      displayName: 'DeepSeek V2 Chat',
-      id: 'deepseek-ai/DeepSeek-V2-Chat',
-      tokens: 32_768,
-    },
-    {
-      description: 'DeepSeek Coder V2 为代码任务设计, 专注于高效的代码生成。',
-      displayName: 'DeepSeek V2 Coder',
-      id: 'deepseek-ai/DeepSeek-Coder-V2-Instruct',
-      tokens: 32_768,
-    },
-    {
-      description: 'DeepSeek 67B 是为高复杂性对话训练的先进模型。',
-      displayName: 'DeepSeek LLM 67B',
-      id: 'deepseek-ai/deepseek-llm-67b-chat',
+      pricing: {
+        currency: 'CNY',
+        input: 1,
+        output: 1,
+      },
       tokens: 32_768,
     },
     {
@@ -137,6 +165,11 @@ const SiliconCloud: ModelProviderCard = {
       displayName: 'Gemma 2 9B',
       enabled: true,
       id: 'google/gemma-2-9b-it',
+      pricing: {
+        currency: 'CNY',
+        input: 0,
+        output: 0,
+      },
       tokens: 8192,
     },
     {
@@ -144,6 +177,11 @@ const SiliconCloud: ModelProviderCard = {
       displayName: 'Gemma 2 27B',
       enabled: true,
       id: 'google/gemma-2-27b-it',
+      pricing: {
+        currency: 'CNY',
+        input: 1.26,
+        output: 1.26,
+      },
       tokens: 8192,
     },
     {
@@ -151,6 +189,11 @@ const SiliconCloud: ModelProviderCard = {
       displayName: 'Llama 3.1 8B',
       enabled: true,
       id: 'meta-llama/Meta-Llama-3.1-8B-Instruct',
+      pricing: {
+        currency: 'CNY',
+        input: 0,
+        output: 0,
+      },
       tokens: 32_768,
     },
     {
@@ -158,6 +201,11 @@ const SiliconCloud: ModelProviderCard = {
       displayName: 'Llama 3.1 70B',
       enabled: true,
       id: 'meta-llama/Meta-Llama-3.1-70B-Instruct',
+      pricing: {
+        currency: 'CNY',
+        input: 4.13,
+        output: 4.13,
+      },
       tokens: 32_768,
     },
     {
@@ -165,24 +213,11 @@ const SiliconCloud: ModelProviderCard = {
       displayName: 'Llama 3.1 405B',
       enabled: true,
       id: 'meta-llama/Meta-Llama-3.1-405B-Instruct',
-      tokens: 32_768,
-    },
-    {
-      description: 'LLaMA 3 支持大容量文本生成和指令解析。',
-      displayName: 'Llama 3 70B',
-      id: 'meta-llama/Meta-Llama-3-70B-Instruct',
-      tokens: 8192,
-    },
-    {
-      description: 'Mistral 7B 指令微调模型针对对话场景进行了优化，可用于文本生成和对话任务。',
-      displayName: 'Mistral 7B',
-      id: 'mistralai/Mistral-7B-Instruct-v0.2',
-      tokens: 32_768,
-    },
-    {
-      description: 'Mixtral 8x7B 模型支持多语言输入和输出，可用于文本生成和对话任务。',
-      displayName: 'Mistral 8x7B',
-      id: 'mistralai/Mixtral-8x7B-Instruct-v0.1',
+      pricing: {
+        currency: 'CNY',
+        input: 21,
+        output: 21,
+      },
       tokens: 32_768,
     },
   ],

package/src/const/fetch.ts

@@ -1,5 +1,6 @@
 export const OPENAI_END_POINT = 'X-openai-end-point';
 export const OPENAI_API_KEY_HEADER_KEY = 'X-openai-api-key';
+export const LOBE_USER_ID = 'X-lobe-user-id';
 
 export const USE_AZURE_OPENAI = 'X-use-azure-openai';
 
@@ -19,9 +20,10 @@ export const getOpenAIAuthFromRequest = (req: Request) => {
   const useAzureStr = req.headers.get(USE_AZURE_OPENAI);
   const apiVersion = req.headers.get(AZURE_OPENAI_API_VERSION);
   const oauthAuthorizedStr = req.headers.get(OAUTH_AUTHORIZED);
+  const userId = req.headers.get(LOBE_USER_ID);
 
   const oauthAuthorized = !!oauthAuthorizedStr;
   const useAzure = !!useAzureStr;
 
-  return { accessCode, apiKey, apiVersion, endpoint, oauthAuthorized, useAzure };
+  return { accessCode, apiKey, apiVersion, endpoint, oauthAuthorized, useAzure, userId };
 };

package/src/libs/agent-runtime/AgentRuntime.ts

@@ -35,6 +35,7 @@ import {
   EmbeddingsPayload,
   ModelProvider,
   TextToImagePayload,
+  TextToSpeechPayload,
 } from './types';
 import { LobeUpstageAI } from './upstage';
 import { LobeZeroOneAI } from './zeroone';
@@ -97,6 +98,9 @@ class AgentRuntime {
   async embeddings(payload: EmbeddingsPayload, options?: EmbeddingsOptions) {
     return this._runtime.embeddings?.(payload, options);
   }
+  async textToSpeech(payload: TextToSpeechPayload, options?: EmbeddingsOptions) {
+    return this._runtime.textToSpeech?.(payload, options);
+  }
 
   /**
    * @description Initialize the runtime with the provider and the options

package/src/libs/agent-runtime/BaseAI.ts

@@ -1,6 +1,5 @@
 import OpenAI from 'openai';
 
-import { TextToImagePayload } from '@/libs/agent-runtime/types/textToImage';
 import { ChatModelCard } from '@/types/llm';
 
 import {
@@ -9,6 +8,9 @@ import {
   EmbeddingItem,
   EmbeddingsOptions,
   EmbeddingsPayload,
+  TextToImagePayload,
+  TextToSpeechOptions,
+  TextToSpeechPayload,
 } from './types';
 
 export interface LobeRuntimeAI {
@@ -20,6 +22,11 @@ export interface LobeRuntimeAI {
   models?(): Promise<any>;
 
   textToImage?: (payload: TextToImagePayload) => Promise<string[]>;
+
+  textToSpeech?: (
+    payload: TextToSpeechPayload,
+    options?: TextToSpeechOptions,
+  ) => Promise<ArrayBuffer>;
 }
 
 export abstract class LobeOpenAICompatibleRuntime {

package/src/libs/agent-runtime/types/index.ts

@@ -1,4 +1,5 @@
 export * from './chat';
 export * from './embeddings';
 export * from './textToImage';
+export * from './tts';
 export * from './type';

package/src/libs/agent-runtime/types/tts.ts

@@ -0,0 +1,14 @@
+export interface TextToSpeechPayload {
+  input: string;
+  model: string;
+  voice: string;
+}
+
+export interface TextToSpeechOptions {
+  headers?: Record<string, any>;
+  signal?: AbortSignal;
+  /**
+   * userId for the embeddings
+   */
+  user?: string;
+}

package/src/libs/agent-runtime/utils/openaiCompatibleFactory/index.ts

@@ -1,7 +1,6 @@
 import OpenAI, { ClientOptions } from 'openai';
 
 import { LOBE_DEFAULT_MODEL_LIST } from '@/config/modelProviders';
-import { TextToImagePayload } from '@/libs/agent-runtime/types/textToImage';
 import { ChatModelCard } from '@/types/llm';
 
 import { LobeRuntimeAI } from '../../BaseAI';
@@ -13,6 +12,9 @@ import {
   EmbeddingItem,
   EmbeddingsOptions,
   EmbeddingsPayload,
+  TextToImagePayload,
+  TextToSpeechOptions,
+  TextToSpeechPayload,
 } from '../../types';
 import { AgentRuntimeError } from '../createError';
 import { debugResponse, debugStream } from '../debugStream';
@@ -253,6 +255,19 @@ export const LobeOpenAICompatibleFactory = <T extends Record<string, any> = any>
       }
     }
 
+    async textToSpeech(payload: TextToSpeechPayload, options?: TextToSpeechOptions) {
+      try {
+        const mp3 = await this.client.audio.speech.create(payload as any, {
+          headers: options?.headers,
+          signal: options?.signal,
+        });
+
+        return mp3.arrayBuffer();
+      } catch (error) {
+        throw this.handleError(error);
+      }
+    }
+
     private handleError(error: any): ChatCompletionErrorPayload {
       let desensitizedEndpoint = this.baseURL;
 

package/src/services/_header.ts

@@ -1,4 +1,9 @@
-import { LOBE_CHAT_ACCESS_CODE, OPENAI_API_KEY_HEADER_KEY, OPENAI_END_POINT } from '@/const/fetch';
+import {
+  LOBE_CHAT_ACCESS_CODE,
+  LOBE_USER_ID,
+  OPENAI_API_KEY_HEADER_KEY,
+  OPENAI_END_POINT,
+} from '@/const/fetch';
 import { useUserStore } from '@/store/user';
 import { keyVaultsConfigSelectors } from '@/store/user/selectors';
 
@@ -8,12 +13,14 @@ import { keyVaultsConfigSelectors } from '@/store/user/selectors';
  */
 // eslint-disable-next-line no-undef
 export const createHeaderWithOpenAI = (header?: HeadersInit): HeadersInit => {
-  const openAIConfig = keyVaultsConfigSelectors.openAIConfig(useUserStore.getState());
+  const state = useUserStore.getState();
+  const openAIConfig = keyVaultsConfigSelectors.openAIConfig(state);
 
   // eslint-disable-next-line no-undef
   return {
     ...header,
-    [LOBE_CHAT_ACCESS_CODE]: keyVaultsConfigSelectors.password(useUserStore.getState()),
+    [LOBE_CHAT_ACCESS_CODE]: keyVaultsConfigSelectors.password(state),
+    [LOBE_USER_ID]: state.user?.id || '',
     [OPENAI_API_KEY_HEADER_KEY]: openAIConfig.apiKey || '',
     [OPENAI_END_POINT]: openAIConfig.baseURL || '',
   };

package/src/services/_url.ts

@@ -1,4 +1,4 @@
-// TODO: 未来所有路由需要全部迁移到 trpc
+// TODO: 未来路由需要迁移到 trpc or /webapi
 
 /* eslint-disable sort-keys-fix/sort-keys-fix */
 import { transform } from 'lodash-es';
@@ -38,9 +38,11 @@ export const API_ENDPOINTS = mapWithBasePath({
   // image
   images: '/api/text-to-image/openai',
 
-  // TTS & STT
-  stt: '/api/openai/stt',
-  tts: '/api/openai/tts',
-  edge: '/api/tts/edge-speech',
-  microsoft: '/api/tts/microsoft-speech',
+  // STT
+  stt: '/webapi/stt/openai',
+
+  // TTS
+  tts: '/webapi/tts/openai',
+  edge: '/webapi/tts/edge',
+  microsoft: '/webapi/tts/microsoft',
 });

package/src/store/file/slices/tts/action.ts

@@ -39,7 +39,7 @@ export const createTTSFileSlice: StateCreator<
     };
     const file = new File([blob], fileName, fileOptions);
 
-    const res = await get().uploadWithProgress({ file });
+    const res = await get().uploadWithProgress({ file, skipCheckFileType: true });
 
     return res?.id;
   },

package/src/store/file/slices/upload/action.ts

@@ -29,6 +29,12 @@ interface UploadWithProgressParams {
           type: 'removeFile';
         },
   ) => void;
+  /**
+   * Optional flag to indicate whether to skip the file type check.
+   * When set to `true`, any file type checks will be bypassed.
+   * Default is `false`, which means file type checks will be performed.
+   */
+  skipCheckFileType?: boolean;
 }
 
 interface UploadWithProgressResult {
@@ -52,8 +58,8 @@ export const createFileUploadSlice: StateCreator<
   [],
   FileUploadAction
 > = (set, get) => ({
-  internal_uploadToClientDB: async ({ file, onStatusUpdate }) => {
-    if (!file.type.startsWith('image')) {
+  internal_uploadToClientDB: async ({ file, onStatusUpdate, skipCheckFileType }) => {
+    if (!skipCheckFileType && !file.type.startsWith('image')) {
       onStatusUpdate?.({ id: file.name, type: 'removeFile' });
       message.info({
         content: t('upload.fileOnlySupportInServerMode', {
@@ -158,11 +164,11 @@ export const createFileUploadSlice: StateCreator<
     return data;
   },
 
-  uploadWithProgress: async ({ file, onStatusUpdate, knowledgeBaseId }) => {
+  uploadWithProgress: async (payload) => {
     const { internal_uploadToServer, internal_uploadToClientDB } = get();
 
-    if (isServerMode) return internal_uploadToServer({ file, knowledgeBaseId, onStatusUpdate });
+    if (isServerMode) return internal_uploadToServer(payload);
 
-    return internal_uploadToClientDB({ file, onStatusUpdate });
+    return internal_uploadToClientDB(payload);
   },
 });