MCP Web Scraping & Markdown Server

Wire clean, LLM-ready web reading into your agent. The RenderKit MCP server exposes the extract primitive — any URL becomes token-aware markdown chunks your agent can consume. The renderkit-mcp server installs in one config block and exposes all three render primitives as tools — your agent screenshots, archives, and reads pages without leaving the conversation.

Add it to your MCP client

{
  "mcpServers": {
    "renderkit": {
      "command": "npx",
      "args": ["renderkit-mcp"],
      "env": { "RENDERKIT_API_KEY": "your-key" }
    }
  }
}

Under the hood — the REST call it wraps

curl https://api.renderkit.tech/v1/extract \
  -H "x-api-key: $RENDERKIT_API_KEY" \
  -H "content-type: application/json" \
  -d '{ "url": "https://example.com", "chunk": true, "chunk_max_tokens": 512 }'

What the agent gets

• extract tool for any MCP client
• Boilerplate-stripped, structure-preserving markdown
• Token-aware chunking for context windows
• Pair with screenshot for visual + text grounding

More

See the MCP server overview, or the underlying screenshot, PDF, and markdown APIs.