跳转到主要内容
适用于 Vercel AI SDK 的 Firecrawl 工具，为 AI 应用提供网页抓取、搜索、浏览和数据提取能力。

​
安装

npm install firecrawl-aisdk ai @ai-sdk/openai
设置环境变量： 
FIRECRAWL_API_KEY=fc-your-key
OPENAI_API_KEY=sk-your-key
这些示例使用的是 OpenAI，但 Firecrawl 工具适用于 Vercel AI SDK 支持的任意提供商，包括 Anthropic、Google、Mistral 等。请参阅支持的提供商完整列表。

​
快速上手

import { generateText } from 'ai';
import { openai } from '@ai-sdk/openai';
import { scrapeTool } from 'firecrawl-aisdk';

const { text } = await generateText({
  model: openai('gpt-5-mini'),
  prompt: '抓取 https://firecrawl.dev 并总结其功能',
  tools: { scrape: scrapeTool },
});

​
可用工具

import {
  scrapeTool,         // Scrape single URL
  searchTool,         // Search the web
  browserTool,        // 交互式浏览器会话
  agentTool,          // Autonomous web agent
  mapTool,            // Discover URLs on a site
  crawlTool,          // Crawl multiple pages
  batchScrapeTool,    // Scrape multiple URLs
  extractTool,        // Extract structured data
  pollTool,           // Poll async jobs
  statusTool,         // Check job status
  cancelTool,         // Cancel jobs
} from 'firecrawl-aisdk';

​
示例

​
爬取

import { generateText } from 'ai';
import { openai } from '@ai-sdk/openai';
import { scrapeTool } from 'firecrawl-aisdk';

const { text } = await generateText({
  model: openai('gpt-5-mini'),
  prompt: '抓取 https://firecrawl.dev 并总结它的作用',
  tools: { scrape: scrapeTool },
});

console.log(text);

import { generateText } from 'ai';
import { openai } from '@ai-sdk/openai';
import { searchTool } from 'firecrawl-aisdk';

const { text } = await generateText({
  model: openai('gpt-5-mini'),
  prompt: '搜索 Firecrawl 并总结你的发现',
  tools: { search: searchTool },
});

console.log(text);

​
映射

import { generateText } from 'ai';
import { openai } from '@ai-sdk/openai';
import { mapTool } from 'firecrawl-aisdk';

const { text } = await generateText({
  model: openai('gpt-5-mini'),
  prompt: 'Map https://docs.firecrawl.dev and list the main sections',
  tools: { map: mapTool },
});

console.log(text);

​
Crawl

异步操作——请包含 pollTool 以检查任务状态。 
import { generateText } from 'ai';
import { openai } from '@ai-sdk/openai';
import { crawlTool, pollTool } from 'firecrawl-aisdk';

const { text } = await generateText({
  model: openai('gpt-5-mini'),
  prompt: '爬取 https://docs.firecrawl.dev（限制 3 页）并总结',
  tools: { crawl: crawlTool, poll: pollTool },
});

console.log(text);

​
批量抓取

异步操作——请使用 pollTool 轮询检查任务状态。 
import { generateText } from 'ai';
import { openai } from '@ai-sdk/openai';
import { batchScrapeTool, pollTool } from 'firecrawl-aisdk';

const { text } = await generateText({
  model: openai('gpt-5-mini'),
  prompt: 'Scrape https://firecrawl.dev and https://docs.firecrawl.dev, then compare',
  tools: { batchScrape: batchScrapeTool, poll: pollTool },
});

console.log(text);

​
提取

异步操作——包含 pollTool 来检查任务状态。 
import { generateText } from 'ai';
import { openai } from '@ai-sdk/openai';
import { extractTool, pollTool } from 'firecrawl-aisdk';

const { text } = await generateText({
  model: openai('gpt-5-mini'),
  prompt: '从 https://firecrawl.dev 提取主要功能',
  tools: { extract: extractTool, poll: pollTool },
});

console.log(text);

​
搜索 + 抓取

import { generateText } from 'ai';
import { openai } from '@ai-sdk/openai';
import { searchTool, scrapeTool } from 'firecrawl-aisdk';

const { text } = await generateText({
  model: openai('gpt-5-mini'),
  prompt: '搜索 Firecrawl,抓取首个结果,并说明其功能',
  tools: { search: searchTool, scrape: scrapeTool },
});

console.log(text);

​
流式

import { streamText } from 'ai';
import { openai } from '@ai-sdk/openai';
import { scrapeTool } from 'firecrawl-aisdk';

const result = streamText({
  model: openai('gpt-5-mini'),
  prompt: '抓取 https://firecrawl.dev 并解释它的功能',
  tools: { scrape: scrapeTool },
});

for await (const chunk of result.textStream) {
  process.stdout.write(chunk);
}

​
浏览器

将 browserTool 与 ToolLoopAgent 搭配使用，以实现交互式网页浏览。该 agent 可以在页面之间跳转、点击元素、填写表单并提取数据。 
import { ToolLoopAgent, stepCountIs } from 'ai';
import { openai } from '@ai-sdk/openai';
import { browserTool } from 'firecrawl-aisdk';

const { text } = await new ToolLoopAgent({
  model: openai('gpt-5-mini'),
  tools: { browserTool },
  stopWhen: stepCountIs(25),
}).generate({
  prompt: 'Go to https://news.ycombinator.com, get the top 3 stories with their titles, points, and links.',
});

console.log(text);
将 browserTool 与 searchTool 结合，适用于从搜索起步并通过交互式浏览继续的工作流。 
import { ToolLoopAgent, stepCountIs } from 'ai';
import { openai } from '@ai-sdk/openai';
import { browserTool, searchTool } from 'firecrawl-aisdk';

const { text } = await new ToolLoopAgent({
  model: openai('gpt-5-mini'),
  tools: { browserTool, searchTool },
  stopWhen: stepCountIs(25),
}).generate({
  prompt: '搜索本周顶级 AI 论文,浏览并总结其核心发现。',
});

console.log(text);

​
Agent

使用 agentTool 实现自动化网页数据采集。该 Agent 能自行搜索、浏览并提取数据。 
import { generateText, stepCountIs } from 'ai';
import { openai } from '@ai-sdk/openai';
import { agentTool, pollTool } from 'firecrawl-aisdk';

const { text } = await generateText({
  model: openai('gpt-5-mini'),
  prompt: 'Find the founders of Firecrawl, their roles, and their backgrounds',
  tools: { agent: agentTool, poll: pollTool },
  stopWhen: stepCountIs(10),
});

console.log(text);