Skip to content
MCP server

acrawl

Rust browser agent that scrapes, debugs and replays crawls, usable as an MCP server

Runs locallyActively maintainedStatus and history

The repository was last updated 7 days ago and is published under MIT.

Install
Pricing
Open source
Platforms
CLI, macOS, Linux
Runs
Locally
GitHub stars
16
Last commit
7 days ago
Language
Rust

Similar MCP servers

About acrawl

acrawl is an open-source MCP server for browser automation. It runs locally on your machine and is started by your AI client.

Acrawl is a single Rust binary with a goal-driven agent. You tell it what you want, for example every book title and price across 50 pages, and it works out pagination and returns JSON. It fetches static pages over HTTP and switches to a real stealth browser only when a page needs JavaScript.

Unlike most browser agents, it can also observe: network requests, console logs, WebSocket frames, performance timings, cookies, storage, code coverage and axe-core accessibility audits are all tools. The README counts 43 tools and 25 LLM providers, including Anthropic, OpenAI, Gemini and local Ollama.

A finished crawl can be saved as a JSON script with loops and conditions, then replayed from cron or CI with no model calls. In MCP mode it exposes its tools to clients such as Claude Code, Cursor and VS Code, and it can fork sub-agents onto separate tabs. It is MIT-licensed, runs locally with your own key and has no telemetry.

Install

Run in a terminal

claude mcp add acrawl -- npx -y https://github.com/Mingye-Lu/AgenticCrawler/releases/download/v0.17.0/acrawl-mcp-linux-x64.mcpb

Add to claude_desktop_config.json

{
  "mcpServers": {
    "acrawl": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/Mingye-Lu/AgenticCrawler/releases/download/v0.17.0/acrawl-mcp-linux-x64.mcpb"
      ]
    }
  }
}
Install in CursorOpens Cursor, which asks you to confirm. Or add it by hand:

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "acrawl": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/Mingye-Lu/AgenticCrawler/releases/download/v0.17.0/acrawl-mcp-linux-x64.mcpb"
      ]
    }
  }
}
Install in VS CodeOpens VS Code, which asks you to confirm. Or add it by hand:

Add to .vscode/mcp.json

{
  "servers": {
    "acrawl": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/Mingye-Lu/AgenticCrawler/releases/download/v0.17.0/acrawl-mcp-linux-x64.mcpb"
      ]
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "acrawl": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/Mingye-Lu/AgenticCrawler/releases/download/v0.17.0/acrawl-mcp-linux-x64.mcpb"
      ]
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "acrawl": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/Mingye-Lu/AgenticCrawler/releases/download/v0.17.0/acrawl-mcp-linux-x64.mcpb"
      ]
    }
  }
}

Using several servers? Build one config file with the MCP config generator.

Tools this server provides (1)

Taken from the project documentation; the running server may differ.

run_goal
Run a goal-driven agent over the browser.

What it does

Goal-driven scraping
Describe fields and let the agent handle pagination.
DevTools tools
Network, console, performance and accessibility inspection.
Replayable scripts
Save a crawl as JSON and rerun without an LLM.
25 model providers
Anthropic, OpenAI, Gemini, Ollama and more.
Parallel tabs
Sub-agents split large jobs without duplicate pages.
Own-browser bridge
An extension lets it work in your logged-in Chrome.

Good for

  • Scrape a product catalogue into JSON
  • Debug failed requests on a local checkout page
  • Run an accessibility audit
  • Schedule a repeat crawl from CI

Pricing

Open source

MIT licence; you bring your own LLM key

acrawl FAQ

Which licence?

MIT.

Does it need Python or Docker?

No. It is one binary; Node.js 20 or newer is needed for the bundled browser.

Can it run without an LLM?

Saved scripts replay with no model calls.

Does it collect telemetry?

The README says no telemetry.

Is acrawl free?

Yes. acrawl is open source, so you can use and self-host it for free.

How do I add acrawl to Claude Code?

Run this command in your terminal: claude mcp add acrawl -- npx -y https://github.com/Mingye-Lu/AgenticCrawler/releases/download/v0.17.0/acrawl-mcp-linux-x64.mcpb

What are the best alternatives to acrawl?

The closest alternatives to acrawl are Browser Use, Scrapling MCP Server, Chrome DevTools MCP, Firecrawl MCP Server and Douyin_TikTok_Download_API.

Where this comes from

  • Description: written from the project's README.
  • Install steps: from the official MCP Registry entry.

How AIPoweredBot checks servers