跳到正文
CrawlForge MCP

No sign-up required

Try CrawlForge tools live. Enter a URL, pick a tool, and see the results instantly.

31tools

3free requests per week

1,000free credits when you sign up

Fetch any URL and get HTML, Markdown, or plain text1 credit

0/3 free requests used this week. Sign up for 1,000 free credits

Output

Ready

Enter a URL and select a tool to get started

Results will appear here

01Start from a preset

Popular examples

Click an example to pre-fill the playground and try it out.

02Three steps

How it works

Three simple steps to extract structured data from any web page.

  1. Step 1

    Enter a URL

    Paste any web page URL or enter a search query.

  2. Step 2

    Pick a tool

    Choose from 31 specialized extraction tools.

  3. Step 3

    Get results

    Instantly receive structured data, ready for your AI pipeline.

03The full rack

All 31 playground tools

A complete toolkit for web data extraction, organized by category. One of them, browser_session, keeps one browser page alive across several calls, so it runs on the API rather than in this one-shot form.
  • 3 creditsBrowser SessionKeep one browser page alive across several calls: open, snapshot, act, readbrowser_sessionAPI only

04From console to code

Use it in your code

Every playground request can be replicated with a single MCP call. Copy and integrate into your project.
Claude Desktop / Cursor
// Add to your MCP config{  "mcpServers": {    "crawlforge": {      "command": "npx",      "args": [        "-y",        "crawlforge-mcp-server"      ]    }  }} // Then ask your AI assistant:// "Extract the main content from//  https://example.com using CrawlForge"
TypeScript / Node.js
// Every tool is a POST to /api/v1/tools/<tool_name>const response = await fetch(  'https://www.crawlforge.dev/api/v1/tools/extract_content',  {    method: 'POST',    headers: {      'X-API-Key': process.env.CRAWLFORGE_API_KEY!,      'Content-Type': 'application/json',    },    body: JSON.stringify({ url: 'https://example.com' }),  },); const result = await response.json(); console.log(result.data.title);console.log(result.data.content);console.log(result.credits_used);      // 2console.log(result.credits_remaining);

05Playground FAQ

Frequently Asked Questions

01Do I need an account to use the playground?

No. You get 3 free requests per week without signing up. Create a free account to unlock 1,000 credits and remove the weekly limit.

02Are playground results the same as the API?

Yes. The playground uses the same CrawlForge MCP tools and returns identical results. Copy any result directly into your application.

03How are credits calculated?

Each tool has a fixed credit cost from 1 to 10 credits per request. Simple extractions cost 1-2 credits, while advanced tools like deep_research cost 10 credits.

04Can I scrape any website?

CrawlForge respects robots.txt and rate limits. Some sites may block automated access. Use stealth_mode (5 credits) for sites that require browser-like behavior.

05What makes CrawlForge different from other scraping tools?

CrawlForge is MCP-native — built from the ground up on the Model Context Protocol. This means your AI agents can use CrawlForge tools directly without REST wrappers or custom integrations.

06How do I integrate CrawlForge into my project?

Add CrawlForge to your MCP config (Claude Desktop, Cursor, Windsurf) or use the API directly with your API key. Check our docs for step-by-step setup guides.

From playground to production

Ready for production?

Sign up for 1,000 free credits and integrate CrawlForge into your AI pipeline.

Get Started FreeView Documentation