Three ways to rip any YouTube transcript: → Website: paste a URL, get the transcript. No signup. → Chrome extension: one click on any YouTube page. → MCP server: feed transcripts directly into Claude, Cursor, or any AI agent. Output in RAG-ready text, structured JSON, or readable Markdown. Multi-strategy extraction (Innertube API → HTML scraping → transcript panel) so it works even when YouTube's API is flaky. Runs locally. No API keys, no accounts, no data collection.

Lightfield — AI-native CRM that builds itself and does work for you

AI-native CRM that builds itself and does work for you

Hey PH! I'm Marsel, product designer in Tokyo. Why I built this: When I start a project, I build a research folder. Text files, docs, anything I can find. But some of the best information is niche and only exists as video. No articles, no blog posts, just some guy explaining it on YouTube. Most transcript tools either push you toward a paid plan or give you timestamped chunks you have to clean up before an AI can use them. So I made a Chrome extension that gives you clean text in one click. Later, working with Claude Code, I noticed it had to write a Python script, install packages, and run the whole thing just to pull a transcript. Extra steps, extra tokens, extra time. So I turned the extension into an MCP server and published it on npm. Now any agent gets clean transcripts in one tool call. Everything runs locally, fully open source, no API key or signup needed. Would love to hear what you think.