@lobehub/chat 1.90.4 → 1.91.1

package/docs/usage/providers/zeroone.mdx

@@ -1,9 +1,8 @@
 ---
 title: Using 01 AI API Key in LobeChat
 description: >-
-  Learn how to integrate and use 01 AI in LobeChat with step-by-step
-  instructions. Obtain an API key, configure 01 AI, and start conversations with
-  AI models.
+  Learn how to integrate and use 01 AI in LobeChat with step-by-step instructions. Obtain an API key, configure 01 AI, and start conversations with AI models.
+
 tags:
   - 01.AI
   - Web UI

package/docs/usage/providers/zeroone.zh-CN.mdx

@@ -1,8 +1,8 @@
 ---
 title: 在 LobeChat 中使用 01.AI 零一万物 API Key
 description: >-
-  学习如何在 LobeChat 中配置并使用 01.AI 零一万物提供的 AI 模型进行对话。获取 API 密钥、填入设置项、选择模型，开始与 AI
-  助手交流。
+  学习如何在 LobeChat 中配置并使用 01.AI 零一万物提供的 AI 模型进行对话。获取 API 密钥、填入设置项、选择模型，开始与 AI 助手交流。
+
 tags:
   - LobeChat
   - 01.AI

package/docs/usage/providers/zhipu.mdx

@@ -1,9 +1,8 @@
 ---
 title: Using Zhipu ChatGLM API Key in LobeChat
 description: >-
-  Learn how to integrate and utilize Zhipu AI models in LobeChat for enhanced
-  conversational experiences. Obtain the API key, configure settings, and start
-  engaging with cognitive intelligence.
+  Learn how to integrate and utilize Zhipu AI models in LobeChat for enhanced conversational experiences. Obtain the API key, configure settings, and start engaging with cognitive intelligence.
+
 tags:
   - Zhipu AI
   - ChatGLM

package/docs/usage/providers/zhipu.zh-CN.mdx

@@ -35,7 +35,7 @@ tags:
   - 打开智谱 AI 并填入获得的 API 密钥
   - 为你的助手选择一个智谱 AI 的模型即可开始对话
 
-  <Image alt={' 选择并使用智谱AI模型 '} inStep src={ 'https://github.com/lobehub/lobe-chat/assets/34400653/b83da559-73d1-4734-87d5-5e22955a9da2' } />
+  <Image alt={' 选择并使用智谱AI模型 '} inStep src={'https://github.com/lobehub/lobe-chat/assets/34400653/b83da559-73d1-4734-87d5-5e22955a9da2'} />
 
   <Callout type={'warning'}>
     在使用过程中你可能需要向 API 服务提供商付费，请参考智谱 AI 的费用政策。

package/docs/usage/providers.mdx

@@ -1,9 +1,8 @@
 ---
 title: Enhancing LobeChat with Multiple Model Providers for AI Conversations
 description: >-
-  Discover how LobeChat offers diverse AI conversation options by supporting
-  multiple model providers, providing flexibility and a wide range of choices
-  for users and developers.
+  Discover how LobeChat offers diverse AI conversation options by supporting multiple model providers, providing flexibility and a wide range of choices for users and developers.
+
 tags:
   - LobeChat
   - AI Conversations

package/docs/usage/start.mdx

@@ -1,9 +1,8 @@
 ---
 title: Get started with LobeChat
 description: >-
-  Explore the exciting features in LobeChat, including Vision Model, TTS & STT,
-  Local LLMs, and Multi AI Providers. Discover more about Agent Market, Plugin
-  System, and Personalization.
+  Explore the exciting features in LobeChat, including Vision Model, TTS & STT, Local LLMs, and Multi AI Providers. Discover more about Agent Market, Plugin System, and Personalization.
+
 tags:
   - Feature Overview
   - Vision Model

package/docs/usage/tools-calling/anthropic.mdx

@@ -1,8 +1,8 @@
 ---
 title: Anthropic Claude 系列 Tools Calling 评测
 description: >-
-  使用 LobeChat 测试 Anthropic Claude 系列模型（Claude 3.5 sonnet / Claude 3 Opus /
-  Claude 3 haiku） 的工具调用（Function Calling）能力，并展现评测结果
+  使用 LobeChat 测试 Anthropic Claude 系列模型（Claude 3.5 sonnet / Claude 3 Opus / Claude 3 haiku） 的工具调用（Function Calling）能力，并展现评测结果
+
 tags:
   - Tools Calling
   - Benchmark

package/docs/usage/tools-calling/anthropic.zh-CN.mdx

@@ -1,8 +1,8 @@
 ---
 title: Anthropic Claude 系列 Tools Calling 评测
 description: >-
-  使用 LobeChat 测试 Anthropic Claude 系列模型（Claude 3.5 sonnet / Claude 3 Opus /
-  Claude 3 haiku） 的工具调用（Function Calling）能力，并展现评测结果
+  使用 LobeChat 测试 Anthropic Claude 系列模型（Claude 3.5 sonnet / Claude 3 Opus / Claude 3 haiku） 的工具调用（Function Calling）能力，并展现评测结果
+
 tags:
   - Tools Calling
   - Benchmark

package/docs/usage/tools-calling/google.mdx

@@ -1,8 +1,8 @@
 ---
 title: Google Gemini 系列 Tool Calling  评测
 description: >-
-  使用 LobeChat 测试 Google Gemini 系列模型（Gemini 1.5 Pro / Gemini 1.5 Flash）
-  的工具调用（Function Calling）能力，并展现评测结果
+  使用 LobeChat 测试 Google Gemini 系列模型（Gemini 1.5 Pro / Gemini 1.5 Flash） 的工具调用（Function Calling）能力，并展现评测结果
+
 tags:
   - Tools Calling
   - Benchmark

package/docs/usage/tools-calling/google.zh-CN.mdx

@@ -1,8 +1,8 @@
 ---
 title: Google Gemini 系列 Tools Calling  评测
 description: >-
-  使用 LobeChat 测试 Google Gemini 系列模型（Gemini 1.5 Pro / Gemini 1.5
-  Flash）的工具调用（Function Calling）能力，并展现评测结果
+  使用 LobeChat 测试 Google Gemini 系列模型（Gemini 1.5 Pro / Gemini 1.5 Flash）的工具调用（Function Calling）能力，并展现评测结果
+
 tags:
   - Tools Calling
   - Benchmark
@@ -21,8 +21,8 @@ Google Gemini 系列模型 Tools Calling 能力一览：
 | Gemini 1.5 Flash | ❌                | ❌           | ❌            | ⛔      | ⛔    |
 
 <Callout type={'important'}>
-  根据我们的的实际测试，强烈建议不要给 Gemini 开启插件，因为目前（截止 2024.07.07）它的 Tools Calling
-  能力实在太烂了。
+  根据我们的的实际测试，强烈建议不要给 Gemini 开启插件，因为目前（截止 2024.07.07）它的 Tools
+  Calling 能力实在太烂了。
 </Callout>
 
 ## Gemini 1.5 Pro

package/docs/usage/tools-calling/groq.zh-CN.mdx

@@ -1,8 +1,8 @@
 ---
 title: Groq Tools Calling
 description: >-
-  了解 Groq 平台模型 Tools Calling的能力一览，包括LLAMA3 70B、LLAMA3
-  8B和Mixtral-8x7B的简单和复杂指令调用情况。
+  了解 Groq 平台模型 Tools Calling的能力一览，包括LLAMA3 70B、LLAMA3 8B和Mixtral-8x7B的简单和复杂指令调用情况。
+
 tags:
   - Groq 平台模型
   - Tools Calling

package/docs/usage/tools-calling/openai.mdx

@@ -1,8 +1,8 @@
 ---
 title: OpenAI GPT 系列 Tools Calling 评测
 description: >-
-  使用 LobeChat 测试 OpenAI GPT 系列模型（GPT 3.5-turbo / GPT-4 /GPT-4o） 的工具调用（Function
-  Calling）能力，并展现评测结果
+  使用 LobeChat 测试 OpenAI GPT 系列模型（GPT 3.5-turbo / GPT-4 /GPT-4o） 的工具调用（Function Calling）能力，并展现评测结果
+
 tags:
   - Tools Calling
   - Benchmark

package/docs/usage/tools-calling/openai.zh-CN.mdx

@@ -1,8 +1,8 @@
 ---
 title: OpenAI GPT 系列 Tools Calling 评测
 description: >-
-  使用 LobeChat 测试 OpenAI GPT 系列模型（GPT 3.5-turbo / GPT-4 /GPT-4o） 的工具调用（Function
-  Calling）能力，并展现评测结果
+  使用 LobeChat 测试 OpenAI GPT 系列模型（GPT 3.5-turbo / GPT-4 /GPT-4o） 的工具调用（Function Calling）能力，并展现评测结果
+
 tags:
   - Tools Calling
   - Benchmark

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@lobehub/chat",
-  "version": "1.90.4",
+  "version": "1.91.1",
   "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/packages/web-crawler/src/crawImpl/exa.ts

@@ -0,0 +1,93 @@
+import { CrawlImpl, CrawlSuccessResult } from '../type';
+import { NetworkConnectionError, PageNotFoundError, TimeoutError } from '../utils/errorType';
+import { DEFAULT_TIMEOUT, withTimeout } from '../utils/withTimeout';
+
+interface ExaResults {
+  author?: string;
+  favicon?: string;
+  id?: string;
+  image?: string;
+  publishedDate?: string;
+  summary?: string;
+  text: string;
+  title: string;
+  url: string;
+}
+
+interface ExaResponse {
+  requestId?: string;
+  results: ExaResults[];
+}
+
+export const exa: CrawlImpl = async (url) => {
+  // Get API key from environment variable
+  const apiKey = process.env.EXA_API_KEY;
+
+  let res: Response;
+
+  try {
+    res = await withTimeout(
+      fetch('https://api.exa.ai/contents', {
+        body: JSON.stringify({
+          livecrawl: 'fallback', // always, fallback
+          text: true,
+          urls: [url],
+        }),
+        headers: {
+          'Content-Type': 'application/json',
+          'x-api-key': !apiKey ? '' : apiKey,
+        },
+        method: 'POST',
+      }),
+      DEFAULT_TIMEOUT,
+    );
+  } catch (e) {
+    const error = e as Error;
+    if (error.message === 'fetch failed') {
+      throw new NetworkConnectionError();
+    }
+
+    if (error instanceof TimeoutError) {
+      throw error;
+    }
+
+    throw e;
+  }
+
+  if (!res.ok) {
+    if (res.status === 404) {
+      throw new PageNotFoundError(res.statusText);
+    }
+
+    throw new Error(`Exa request failed with status ${res.status}: ${res.statusText}`);
+  }
+
+  try {
+    const data = (await res.json()) as ExaResponse;
+
+    if (!data.results || data.results.length === 0) {
+      console.warn( 'Exa API returned no results for URL:', url )
+      return
+    }
+
+    const firstResult = data.results[0];
+
+    // Check if content is empty or too short
+    if (!firstResult.text || firstResult.text.length < 100) {
+      return;
+    }
+
+    return {
+      content: firstResult.text,
+      contentType: 'text',
+      length: firstResult.text.length,
+      siteName: new URL(url).hostname,
+      title: firstResult.title,
+      url: firstResult.url || url,
+    } satisfies CrawlSuccessResult;
+  } catch (error) {
+    console.error(error);
+  }
+
+  return;
+};

package/packages/web-crawler/src/crawImpl/firecrawl.ts

@@ -0,0 +1,97 @@
+import { CrawlImpl, CrawlSuccessResult } from '../type';
+import { NetworkConnectionError, PageNotFoundError, TimeoutError } from '../utils/errorType';
+import { DEFAULT_TIMEOUT, withTimeout } from '../utils/withTimeout';
+
+interface FirecrawlMetadata {
+  description: string;
+  keywords: string;
+  language: string;
+  ogDescription?: string;
+  ogImage?: string;
+  ogLocaleAlternate?: string[];
+  ogSiteName?: string;
+  ogTitle?: string;
+  ogUrl?: string;
+  robots: string;
+  statusCode: number;
+  sourceURL: string;
+  title: string;
+}
+
+interface FirecrawlResults {
+  html?: string;
+  markdown?: string;
+  metadata: FirecrawlMetadata;
+}
+
+interface FirecrawlResponse {
+  success: boolean;
+  data: FirecrawlResults;
+}
+
+export const firecrawl: CrawlImpl = async (url) => {
+  // Get API key from environment variable
+  const apiKey = process.env.FIRECRAWL_API_KEY;
+  const baseUrl = process.env.FIRECRAWL_URL || 'https://api.firecrawl.dev/v1';
+
+  let res: Response;
+
+  try {
+    res = await withTimeout(
+      fetch(`${baseUrl}/scrape`, {
+        body: JSON.stringify({
+          formats: ["markdown"], // ["markdown", "html"]
+          url,
+        }),
+        headers: {
+          'Authorization': !apiKey ? '' : `Bearer ${apiKey}`,
+          'Content-Type': 'application/json',
+        },
+        method: 'POST',
+      }),
+      DEFAULT_TIMEOUT,
+    );
+  } catch (e) {
+    const error = e as Error;
+    if (error.message === 'fetch failed') {
+      throw new NetworkConnectionError();
+    }
+
+    if (error instanceof TimeoutError) {
+      throw error;
+    }
+
+    throw e;
+  }
+
+  if (!res.ok) {
+    if (res.status === 404) {
+      throw new PageNotFoundError(res.statusText);
+    }
+
+    throw new Error(`Firecrawl request failed with status ${res.status}: ${res.statusText}`);
+  }
+
+  try {
+    const data = (await res.json()) as FirecrawlResponse;
+
+    // Check if content is empty or too short
+    if (!data.data.markdown || data.data.markdown.length < 100) {
+      return;
+    }
+
+    return {
+      content: data.data.markdown,
+      contentType: 'text',
+      description: data.data.metadata.description,
+      length: data.data.markdown.length,
+      siteName: new URL(url).hostname,
+      title: data.data.metadata.title,
+      url: url,
+    } satisfies CrawlSuccessResult;
+  } catch (error) {
+    console.error(error);
+  }
+
+  return;
+};

package/packages/web-crawler/src/crawImpl/index.ts

@@ -1,13 +1,19 @@
 import { browserless } from './browserless';
+import { exa } from './exa';
+import { firecrawl } from './firecrawl';
 import { jina } from './jina';
 import { naive } from './naive';
 import { search1api } from './search1api';
+import { tavily } from './tavily';
 
 export const crawlImpls = {
   browserless,
+  exa,
+  firecrawl,
   jina,
   naive,
   search1api,
+  tavily,
 };
 
 export type CrawlImplType = keyof typeof crawlImpls;

package/packages/web-crawler/src/crawImpl/tavily.ts

@@ -0,0 +1,94 @@
+import { CrawlImpl, CrawlSuccessResult } from '../type';
+import { NetworkConnectionError, PageNotFoundError, TimeoutError } from '../utils/errorType';
+import { DEFAULT_TIMEOUT, withTimeout } from '../utils/withTimeout';
+
+interface TavilyResults {
+  images?: string[];
+  raw_content: string;
+  url: string;
+}
+
+interface TavilyFailedResults {
+  error?: string;
+  url: string;
+}
+
+interface TavilyResponse {
+  base_url: string;
+  failed_results?: TavilyFailedResults[];
+  response_time: number;
+  results: TavilyResults[];
+}
+
+export const tavily: CrawlImpl = async (url) => {
+  // Get API key from environment variable
+  const apiKey = process.env.TAVILY_API_KEY;
+
+  let res: Response;
+
+  try {
+    res = await withTimeout(
+      fetch('https://api.tavily.com/extract', {
+        body: JSON.stringify({
+          extract_depth: process.env.TAVILY_EXTRACT_DEPTH || 'basic', // basic or advanced
+          include_images: false,
+          urls: url,
+        }),
+        headers: {
+          'Authorization': !apiKey ? '' : `Bearer ${apiKey}`,
+          'Content-Type': 'application/json',
+        },
+        method: 'POST',
+      }),
+      DEFAULT_TIMEOUT,
+    );
+  } catch (e) {
+    const error = e as Error;
+    if (error.message === 'fetch failed') {
+      throw new NetworkConnectionError();
+    }
+
+    if (error instanceof TimeoutError) {
+      throw error;
+    }
+
+    throw e;
+  }
+
+  if (!res.ok) {
+    if (res.status === 404) {
+      throw new PageNotFoundError(res.statusText);
+    }
+
+    throw new Error(`Tavily request failed with status ${res.status}: ${res.statusText}`);
+  }
+
+  try {
+    const data = (await res.json()) as TavilyResponse;
+
+    if (!data.results || data.results.length === 0) {
+      console.warn( 'Tavily API returned no results for URL:', url )
+      return
+    }
+
+    const firstResult = data.results[0];
+
+    // Check if content is empty or too short
+    if (!firstResult.raw_content || firstResult.raw_content.length < 100) {
+      return;
+    }
+
+    return {
+      content: firstResult.raw_content,
+      contentType: 'text',
+      length: firstResult.raw_content.length,
+      siteName: new URL(url).hostname,
+      title: new URL(url).hostname,
+      url: firstResult.url || url,
+    } satisfies CrawlSuccessResult;
+  } catch (error) {
+    console.error(error);
+  }
+
+  return;
+};

package/src/app/(backend)/webapi/user/avatar/[id]/[image]/route.ts

@@ -2,7 +2,7 @@ import { UserService } from '@/server/services/user';
 
 export const runtime = 'nodejs';
 
-type Params = Promise<{ id: string , image: string }>;
+type Params = Promise<{ id: string; image: string }>;
 
 // 扩展名到内容类型的映射
 const CONTENT_TYPE_MAP: Record<string, string> = {

package/src/app/[variants]/(main)/repos/[id]/_layout/Mobile.tsx

@@ -22,13 +22,13 @@ const Layout = memo<LayoutProps>(({ children }) => {
 
   return (
     <Flexbox
-        className={styles.main}
-        height="100%"
-        style={showMobileWorkspace ? { display: 'none' } : undefined}
-        width="100%"
-      >
-        {children}
-      </Flexbox>
+      className={styles.main}
+      height="100%"
+      style={showMobileWorkspace ? { display: 'none' } : undefined}
+      width="100%"
+    >
+      {children}
+    </Flexbox>
   );
 });
 

package/src/config/aiModels/deepseek.ts

@@ -1,4 +1,5 @@
 import { AIChatModelCard } from '@/types/aiModel';
+
 // https://api-docs.deepseek.com/zh-cn/quick_start/pricing
 const deepseekChatModels: AIChatModelCard[] = [
   {

package/src/config/aiModels/hunyuan.ts

@@ -1,4 +1,5 @@
 import { AIChatModelCard } from '@/types/aiModel';
+
 // https://cloud.tencent.com/document/product/1729/104753
 const hunyuanChatModels: AIChatModelCard[] = [
   {

package/src/config/aiModels/mistral.ts

@@ -9,8 +9,7 @@ const mistralChatModels: AIChatModelCard[] = [
       functionCall: true,
     },
     contextWindowTokens: 128_000,
-    description:
-      'Mistral Medium 3 以 8 倍的成本提供最先进的性能，并从根本上简化了企业部署。',
+    description: 'Mistral Medium 3 以 8 倍的成本提供最先进的性能，并从根本上简化了企业部署。',
     displayName: 'Mistral Medium 3',
     enabled: true,
     id: 'mistral-medium-latest',

package/src/config/aiModels/modelscope.ts

@@ -6,10 +6,10 @@ const modelscopeChatModels: AIChatModelCard[] = [
       functionCall: true,
     },
     contextWindowTokens: 131_072,
-    description: 'DeepSeek-V3是DeepSeek第三代模型，在多项基准测试中表现优异。',
-    displayName: 'DeepSeek-V3-0324',
+    description: 'DeepSeek R1 通过利用增加的计算资源和在后训练过程中引入算法优化机制，显著提高了其推理和推断能力的深度。该模型在各种基准评估中表现出色，包括数学、编程和一般逻辑方面。其整体性能现已接近领先模型，如 O3 和 Gemini 2.5 Pro。',
+    displayName: 'DeepSeek-R1-0528',
     enabled: true,
-    id: 'deepseek-ai/DeepSeek-V3-0324',
+    id: 'deepseek-ai/DeepSeek-R1-0528',
     type: 'chat',
   },
   {