Convert lecture files into notes that can be checked against source frames
At a glance
A student built Lore Engine to turn lecture videos, transcripts, and PDFs into illustrated Markdown study notes. This workflow is listed under studying, lecture-notes, document-processing.
- Agents / tools
- Lore Engine, Gemini, uv
SourceBuilding a tool for these workflows? See weekly rates.
What they did
1) Install uv and clone https://github.com/Slydite/lore-engine.git. Enter the repository and run `uv sync`. Obtain a Gemini API key and put `GEMINI_API_KEY_1=YOUR_KEY` in a project-root `.env` file. Windows users may also need FFmpeg. 2) Prepare a lecture PDF or a video with an accompanying `.srt` transcript; video transcription is a separate prerequisite. 3) Run `uv run python src/main.py --path /path/to/lecture.mp4`, substituting your file, or use `--batch-path /path/to/lectures/` for a folder. 4) Open the generated Markdown and image files in the output directory. During study, check explanations against the included images: video-frame filenames carry timestamps and PDF-image filenames carry page numbers.
How it went
The repository includes illustrated notes generated from a one-hour MIT Cognitive Robotics lecture as a worked output example.
Worth knowing
Content is sent to Gemini even though the program runs locally. The README identifies a legacy Google SDK; the author's headline time-saving claim is not a measured result for a reader's files.
