Edit a sound recording with AI
Take the sound out of a video, join compatible recordings or save an excerpt for a podcast. The tools work on files in your own workspace.
To edit a sound recording with AI, start with a concrete deliverable such as the short introduction for a podcast episode. Give the agent the take you want and the approximate start and end. Listen to the cut before dropping it into the episode, particularly if the source was compressed.
The audio skill uses FFmpeg for local file work. It can extract the soundtrack of a video, adjust loudness, change the sample rate and join takes. A join copies matching streams, so a field recorder and a camera may need to be converted to the same settings first.
What it does
- Extract a soundtrack to review a recorded conversation without the picture.
- Cut a short teaser using the start and end of the sentence you want.
- Join matching session files in the order they should play.
- Bring a take to a more consistent loudness and keep the WAV result for your editing session.
- Add the built-in one-second fade-in to the beginning of a music cue.
Install and use
- Install the skill, uv, Python 3.11 or later and FFmpeg with ffprobe, then grant access to the source file.
- Ask, “Extract the interview from this video and save a separate WAV for the editor.”
- Listen to the beginning, the ending and any join. Keep the original take available while you choose the final cut.
FAQ
Source
Made by Timelapse Tech. Original skill: audio, published on GitHub under the MIT licence.
View on GitHub