Wraps the Thunderbit Open API and exposes seven tools for web scraping inside Claude, Cursor, and other MCP clients. You can distill pages to clean Markdown (1 credit), extract structured data by passing a JSON Schema (20 credits), or ask the API to suggest fields when you don't know the page structure yet. Batch operations let you queue up to 100 URLs at once, and status polling is free. Setup is a one-liner in your MCP config with an API key from app.thunderbit.com. Useful when you need web content in a shape LLMs can digest or want to pull specific fields from pages without writing scrapers.
claude mcp add --transport stdio thunderbit-open-thunderbit-mcp-server uvx thunderbit-mcp-server