What it is
Audio/video transcoding is an official skill in the Skill Library: add cloud audio/video transcoding to your product—convert video asynchronously to standard MP4, or convert audio-only input asynchronously to MP3.Where to find it
- Open a project → Build → Skill Library.
- Find Audio/video transcoding.
- Click Use and follow the in-product guide to enable it.
Capabilities
Standard playback formats
- Video: output MP4 + H.264 + AAC
- Audio only: output MP3
Async jobs
Submit a job and get a task ID immediately, then poll status—without occupying Edge Function execution time.Reuse existing file storage
Upload media to the session’s existing File storage. The platform writes outputs back to the same session directory and does not create a separate bucket for transcoding.Billed on success
- Video: billed by duration and resolution
- Audio only: billed by duration
- Failed jobs do not charge the transcoding fee (other related costs such as storage still follow the bill)
Integration flow
- After enabling audio/video transcoding, upload the user’s audio or video to file storage and obtain an accessible URL.
- Submit the transcoding job from an Edge Function (save the returned task ID).
- Poll job status; when done, play or download the media from the output URL in the same hosted file storage.
Exact API paths, supported source formats, and unit prices follow in-product configuration and billing. Users typically complete these steps through chat or an already wired backend.
FAQ
Do failed jobs still charge?
Do failed jobs still charge?
Failed jobs do not charge the transcoding fee; successful jobs are billed per the rules. Storage and other costs still follow the bill.
Where is the output stored?
Where is the output stored?
Written back to the existing file-storage directory for the session used at submit time—no separate bucket. Preview, download, or copy the URL in file storage.
Do I need to stay on the page?
Do I need to stay on the page?
No. After submit you can poll with the task ID; the Edge Function is not held for the full transcoding duration.

