Rust MCP server that gives AI agents reliable web access. Bypasses Cloudflare and bot protection using Chrome-level TLS fingerprinting, no headless browser needed. 10 tools for scraping, crawling, search, structured extraction, summarization, and more. Output is clean markdown optimized for LLMs with 67% fewer tokens than raw HTML. 8 of 10 tools work locally without an API key.
Overview
webclaw MCP Server
10 tools for web extraction:
- scrape — extract clean content from any URL
- crawl — recursive site crawling with depth control
- search — web search with result scraping
- extract — structured JSON extraction using LLM
- summarize — page summaries
- brand — extract colors, fonts, logos
- diff — track content changes between snapshots
- map — discover URLs from sitemaps
- batch — parallel multi URL extraction
- research — deep multi source analysis
Setup
npx create-webclaw
Auto detects Claude Desktop, Claude Code, Cursor, Windsurf, Codex, OpenCode and writes the correct config for each one.
Manual config
{
"mcpServers": {
"webclaw": {
"command": "webclaw-mcp",
"args": []
}
}
}
Install the binary
brew tap 0xMassi/webclaw && brew install webclaw
Or download from GitHub releases: https://github.com/0xMassi/webclaw/releases
MIT licensed. https://github.com/0xMassi/webclaw
Server Config
{
"mcpServers": {
"webclaw": {
"command": "webclaw-mcp",
"args": []
}
}
}