Pull audio from a Descript export for $0.01 with Sume Audio detach

Descript Creator exports 4K video. If you need only the sound, Sume Audio detach turns one hosted video into a wav or mp3 for a flat $0.01 per job.

5 min readSume
All posts

If a finished video exists and you only need its soundtrack as a file, Sume Audio detach does that for $0.01 a job, and no Descript plan is needed for that step. Descript's pricing page describes text-based editing with media hours and AI credits per tier; its Creator plan lists 30 hours of media and 4K export at $24 a month. Pulling audio out of an already exported file is a smaller task than that, and it is a good fit for a flat-priced job.

Note the direction of the workflow. Descript is where people edit; Sume Audio detach is a post-process on a hosted clip.

Descript plans and Sume audio steps, read 2026-10-08
ItemDescriptSume
TaskEdit video and audio by editing a transcriptDetach the audio track of one hosted video into a wav or mp3
PriceHobbyist $16, Creator $24, Business $50 a month$0.01 per job (Audio detach)
Allowance10, 30 or 40 media hours; 400, 800 or 1,500 AI creditsNone
Next stepExportConcat or split with Timeline audio, $0.01 per job

How the detach job works

Audio detach takes one clip that your workspace already hosts on media.sume.com and returns a durable audio artifact, wav or mp3. The price is a flat $0.01 per job and uses worker ffmpeg, not a provider model, so there is no per-minute surprise. You import the file first with a media import, send an Idempotency-Key, and poll the job envelope for the result.

The result is a Sume-hosted file you can use as the audio spine of a Timeline render, or hand to a transcription or dubbing step elsewhere.

A concrete case is a batch. Suppose you have 50 finished clips and you need the soundtrack of each for a transcription or a review pass. At $0.01 a job that is $0.50 in total, and each job is independent, so you can submit them all with distinct Idempotency-Key values and poll or use a webhook for the results. The one constraint to remember is that the source must already be a hosted clip in your workspace, so files from an editor have to be imported before the detach call. Nothing in this flow reads or changes your Descript project; it only works on the exported video file.

Joining and splitting afterwards

Timeline audio concatenates or splits Sume-hosted audio into durable files at a flat $0.01 per job. It is the right place to merge a narration and a pad of silence into a reusable spine. Use audio.parts[] inside a Timeline render only when the sliced audio is needed inside that render alone.

When the spine is ready, Timeline 1.0 costs $0.10 per output minute. A 90-second render rounds up to 2 minutes and costs $0.20; the audio steps before it cost $0.01 each. A whole detach, join and render for a 90-second piece is therefore $0.01 + $0.01 + $0.20 = $0.22.

  • Detach: one hosted video in, one wav or mp3 out.
  • Join or split: Timeline audio, still $0.01.
  • Render: Timeline 1.0, $0.10 per ceil output minute.

When to stay in the editor

If you are already inside Descript, export the audio there; it is part of the plan you pay for. The flat-priced job pays off when the video comes from somewhere else, such as a generated clip, and a script needs the sound without opening an editor.

Sources

Related posts

More in Comparisons

All Comparisons posts

Written by Sume