acrawl — Claude Code MCP server: what it does, how to install
Automates browser tasks, inspects network calls, checks console logs, and crawls complex sites. Useful when an agent needs full DevTools visibility.
Open on GitHubWhat acrawl does
Built as a standalone Rust binary, the server connects via acrawl mcp. It exposes 38 tools plus a goal-driven sub-agent to the model. Rather than just clicking elements, it surfaces DevTools capabilities: request timing inspection, WebSocket logging, TTFB metrics, and axe-core accessibility checks.
The engine serves static pages over HTTP for speed. When detecting frontend markers from React, Vue, or Next.js, it switches to a headless browser automatically. You can also run deterministic JSON routines locally, executing loops and branches without wasting model tokens.
Browser automation relies on a local Chromium download: without it, the agent falls back to pure HTTP fetching. Multi-tab crawling runs parallel sub-agents with a shared URL registry to prevent duplicate visits.
What it helps with
- inspect HTTP requests, timings, and console errors
- extract content from pages built with client-side frameworks
- run automated axe-core accessibility audits
- execute deterministic navigation loops without spending LLM tokens
Example prompts for the agent
- Run
run_goalin acrawl to locate pricing details across the product catalog - Use
acrawl mcpto measure TTFB and capture console warnings for the landing page - Audit page accessibility with axe-core and list WCAG issues
How to install acrawl
Install
The MCP registry has no launch command for it: see the repository README
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about acrawl
How do I install acrawl for Claude Code?
Compile the binary with cargo build --release and add acrawl mcp to your MCP configuration. The program runs natively without Python or Node.js.
What does acrawl do?
It provides tools for automated browser control, network inspection, performance profiling, and autonomous goal navigation via run_goal.
What external dependencies does acrawl require?
Running browser sessions requires a Chromium download. Static pages can still be fetched directly over HTTP without a headless browser.
