Acrawl
Give coding agents a stealth, single-binary browser crawler with 17 tools and a goal-driven run_goal loop for research and extraction tasks.
Overview
acrawl is a MCP server for the Build phase that provides 17 browser tools and a goal-driven run_goal agent for autonomous stealth web crawling.
What is this MCP server?
- 17 dedicated browser MCP tools plus a goal-driven run_goal agent mode
- Shipped as a single MCPB binary (stdio) for linux-x64 v0.9.0
- Stealth-oriented crawling for sites that block naive bots
- Autonomous navigation aimed at extraction and multi-step web goals
- stdio transport fits local Claude Code and Cursor MCP configs
- 17 browser MCP tools documented in server description
- Goal-driven run_goal agent mode included
- Single-binary MCPB distribution (stdio), version 0.9.0
Community signal: 5 GitHub stars.
What problem does it solve?
Your agent cannot reliably complete multi-step browser tasks on modern sites using plain HTTP or manual copy-paste.
Who is it for?
Solo builders automating research, live UI checks, and goal-based web extraction inside Claude Code or Cursor.
Skip if: Large-scale production scrapers, legally sensitive data harvests without review, or teams that forbid local browser automation.
What do I get? / Deliverables
You register one stdio MCP binary and delegate structured browsing, extraction, and goal pursuit to the agent from the IDE.
- Agent-driven page navigation and extraction via 17 browser tools
- Goal-completed crawl sessions through run_goal
- Local stealth browsing without a separate automation repo
Recommended MCP Servers
Journey fit
Browser automation MCPs are build-phase agent infrastructure— they extend what Claude/Cursor can do on the live web during product and integration work. Agent-tooling is the canonical shelf for MCP servers that add autonomous browser capabilities to the coding agent stack.
How it compares
Local stealth browser automation MCP, not a hosted SEO scorer or static API wrapper.
Common Questions / FAQ
Who is Acrawl for?
Developers using MCP-enabled agents who need autonomous browser control, many fine-grained tools, and a run_goal mode for multi-step web tasks.
When should I use Acrawl?
Use it while building or validating integrations—competitor pages, docs behind client-side rendering, or repeatable agent-led web research.
How do I add Acrawl to my agent?
Download the Acrawl MCPB release (e.g. v0.9.0 linux-x64 from AgenticCrawler GitHub), configure stdio transport in your MCP settings pointing at the binary, then call browser tools or run_goal from your agent.