Turn YouTube Playlists Into Structured Markdown Books

What it does

This PyQt5 desktop app takes a YouTube playlist or single video URL, pulls every available transcript, and reformats the whole pile into a single markdown file. It chunks transcripts into configurable pieces (default around 3,000 words) and sends them sequentially to the Google Gemini API, passing previously refined text along as context to keep the final output coherent. You pick the output language and a formatting style, and it returns a book-like file ready for Obsidian or NotebookLM.

The interesting bit

Instead of choking the API with a wall of text, the app uses a rolling refinement pipeline: each chunk is processed with the prior chunk’s polished output as context, so a 50-video playlist still reads like one continuous document. The five built-in styles—especially the Educational mode with inline jargon definitions and the foldable Q&A format—show this was built by someone who actually wanted to study the output, not just dump it.

Key highlights

• Batch-processes playlists or single videos through a simple GUI
• Five output styles: detailed, summary, educational glossary, narrative rewrite, and foldable Q&A
• Configurable chunk size to trade API call count against detail fidelity
• Output language is decoupled from the source video language
• Exports clean markdown aimed at note-taking apps and knowledge bases

Verdict

Grab it if you are hoarding educational YouTube content for offline reading or feeding into a second brain. Skip it if you only need the occasional one-off summary; this tool is built for volume.