lectures-digitizer

Converts lecture recordings into structured notes via ffmpeg chunking and Gemini Gem transcription.

Updated Nov 21, 2025
One-click install
npx skills add https://github.com/alatyshau/duet --skill lectures-digitizer-alatyshau
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: lectures-digitizer
Source: https://github.com/alatyshau/duet/tree/main/packages/skills/tools/lectures-digitizer
Command: npx skills add https://github.com/alatyshau/duet --skill lectures-digitizer-alatyshau

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires ffmpeg.

What problem does it solve? Long lecture recordings are hard to turn into usable study notes. This Skill automates the tedious parts—audio conversion, chunking, and final assembly—while coordinating a human-in-the-loop workflow with a Gemini Gem that produces structured summaries of each chunk. ## Core Features & Use Cases - Audio Preparation: Converts any media file to 16kHz mono MP3 with ffmpeg and splits it into 20-minute chunks with 30-second overlap, with caching to skip repeated work. - Gemini Gem Coordination: Prepares chunk files and placeholder digest files, then guides the user through uploading chunks to a Gemini Gem and pasting results back. - Digest Assembly: Merges per-chunk summaries into a single digest.md, removing overlap duplicates and unifying sections like key ideas, terms, and examples. - Use Case: You recorded a 90-minute university lecture. Run the Skill, upload the generated chunks to your Gemini Gem, paste the summaries back, and receive one clean, structured lecture note document. ## Quick Start Ask the agent to digitize a lecture by providing the path to your recording, for example: digitize the lecture at /recordings/physics-lecture-01.mp4.

Frequently Asked Questions about lectures-digitizer

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I convert a lecture recording into structured notes?

Provide the path to your media file and the Skill converts it to MP3 with ffmpeg, splits it into 20-minute chunks, and prepares digest files. You upload the chunks to a Gemini Gem, paste the summaries back, and the Skill assembles a final digest.md.

How to split long audio files into chunks with ffmpeg?

The Skill uses ffmpeg with -ss and -to parameters to cut the converted MP3 into 20-minute chunks with 30-second overlap. If the audio is under 20 minutes, it is copied as a single chunk without splitting.

Does this lecture digitization require a Gemini account?

Yes, summarization happens through a Gemini Gem on gemini.google.com, which requires a Google account. The Skill includes instructions for creating the LecturesDigitizer Gem using a provided instruction file.

What happens if ffmpeg is not installed?

The Skill checks for ffmpeg in PATH during initialization and reports an error if it is missing. On macOS it suggests installing it with brew install ffmpeg before continuing.

Can I rerun the digitization without reprocessing the audio?

Yes, the converted audio.mp3 and the chunks directory are cached between runs, so those steps are skipped on repeat executions. Only the digest directory is regenerated on each new run.