Video to transcript cost: audio detach plus STT, by minutes

Getting a transcript from a video on Sume costs one cent for the detach plus one cent per audio minute for speech-to-text, so a five-minute video is $0.06. The audio detach docs list a public rate of $0.01 per job, and the pricing page lists Sume STT 1.0 at $0.01 per audio minute with a 10-minute maximum. Both were read on 2026-10-03.

What does it cost by video length?

Detach is a flat fee, so its share shrinks as the video grows.

Is there a one-call shortcut?

Yes, for a clip already on Sume. Video inspect with transcribe: true runs STT on the clip's audio and returns text, words[] and optional sentence segments[]; its per-minute rate is added to the inspect reservation. Choose detach plus STT when you want the audio file itself, for example to drive a lip-sync job or to join with other audio, and inspect when you only want words.

What can go wrong?

• A source with no audio track fails with detach_source_has_no_audio; check probe.has_audio first with inspect and frames: false.
• Detach caps the source at 1800 seconds and the output at 900, and STT takes 10 minutes a request, so a 20-minute video needs windows (see the two-hour post in this series).
• The video must already be on your workspace's media.sume.com host. Off-host URLs are refused at admit.