PinRAG MCP server

PinRAG is a helper that lets your AI read your own study materials and answer questions about them. You feed it PDFs, YouTube videos, GitHub repos, Discord chats, or web docs, and then you just ask. It is handy if you learn from many scattered sources and want one place to ask questions.

What this MCP server does

You tell your AI to add some material, like a PDF path or a YouTube link. The AI passes that to PinRAG, which reads the file or link and stores it in a searchable index on your computer. Later, when you ask a question, the AI sends it to PinRAG, which finds the most relevant parts of your material. PinRAG hands those parts back to the AI, and the AI writes an answer for you with citations pointing to the page, timestamp, file, or URL.

Click to zoom

What you can do with it

• Index a PDF, a folder of text files, or a Discord export
• Index a YouTube video or playlist and ask what it covers
• Index a GitHub repo and ask about the code
• Index a documentation website and ask about its features
• Tag documents so you can search only one topic
• List everything you have indexed so far
• Remove a document you no longer need

Try asking your AI

• “Add /Users/me/books/amiga-book.pdf with tag AMIGA”
• “Index https://youtu.be/xyz and tell me what it says about memory”
• “Index https://github.com/owner/repo and explain how the config works”
• “What did the PDF say about interrupts on page 40 to 60?”

What it gives back to you

You get a written answer in the chat, based on your own material. Each answer includes citations, like a PDF page number, a YouTube timestamp such as t. 1:23, a file name, a chunk label for GitHub, or a source URL for web docs. When you index or list, you get short confirmations and a list of document IDs and chunk counts.

Before you start

Who it's for

People who learn from many sources at once, like developers, students, and researchers who want to ask questions across PDFs, videos, repos, and chats.