Lecture Notes Skill

Transform lecture materials into structured, exam-ready Markdown notes for Obsidian.

Audio and Video Inputs

When the source material is an audio or video file, always create a local transcript first and use that transcript as the primary input for note generation.

Required Workflow

1. If the input is audio, transcribe it with OpenRouter using openai/gpt-4o-audio-mini and the input_audio format.
2. If the input is video, first extract a separate audio file with ffmpeg, then transcribe that extracted audio with OpenRouter.
3. Save the transcript locally as a plain text file and use that saved transcript for the rest of the lecture-notes workflow.
4. If useful, also save the raw JSON response locally, but the .txt transcript is the canonical note-generation source.

OpenRouter Audio Transcription

Before using audio or video transcription, make sure OPENROUTER_API_KEY is defined in the environment.

Use OpenRouter's chat completions API with multimodal input_audio. The prompt should explicitly ask for a verbatim transcript with minimal cleanup.