LLM-ready output
Markdown, JSON, text, and an LLM format that removes page markup and boilerplate.
The platform
Webclaw is one Rust extraction engine, served as a hosted Cloud API, a single-binary CLI, and an MCP server for AI agents. Same clean, LLM-ready output everywhere, with SDKs for Python, TypeScript and Go.
Under the hood
The API, CLI, and MCP server share the core extraction library. The hosted API also manages fetching, caching, authentication, and billing, so supported behavior differs between deployments.
Markdown, JSON, text, and an LLM format that removes page markup and boilerplate.
Automatic handling for supported protected and dynamic pages, with errors when the source cannot be accessed.
Fetch static page content directly and request JavaScript rendering when needed. Timing depends on the page and processing.
Markdown, text, json, llm, links, rawHtml, attributes, query, and extract. Screenshots are a separate option.
A single fast engine under every surface, with no heavy infrastructure to babysit.
AGPL-3.0. Run the API, CLI and MCP server on your own hardware without hosted API credits.
Type-safe clients over the same REST API, so you can call every endpoint in the language you already ship in.
Cancel anytime. API, CLI, or MCP, the engine is the same.