Superfetch

Strips HTML down to clean Markdown so your LLM can focus on content instead of navigation bars and ad scripts. Handles any public URL and recognizes GitHub, GitLab, Bitbucket, and Gist links to fetch raw content directly. Returns metadata like title, URL, and content size alongside the converted text. Larger or slower pages can run as async tasks with progress updates instead of blocking your session. The HTTP mode adds bearer auth, rate limiting, and TLS if you need it exposed as an endpoint. Useful when you're scraping documentation, analyzing articles, or pulling repo files into context without the surrounding page chrome.