What it does
audio-tldr is a skill that turns videos, audio and podcasts into 3 to 7 key takeaways and a summary. Transcription runs locally with whisper, and each result is cached by content hash, so the same source is not transcribed again unless you force it. You can ask for a different angle later and get a new digest from the cache.
Use cases
- 01Key points from a long podcast episode
- 02Summary of a recorded webinar or customer call
- 03A second summary from the same cached transcript