$50 of audio on Sume: tracks, minutes and characters it buys
At Sume list rates $50 buys 400 music tracks, 5,000 STT minutes, about a million TTS characters, 250 caption jobs, or 266 seconds of Fabric 720p. One table.
At Sume's list rates, $50 buys 400 music tracks, 5,000 minutes of speech-to-text, about a million characters of text to speech, 250 caption jobs, or 266 seconds of Fabric 720p lip-sync, if you spend it all on one thing. Rates are from the pricing page, the captions docs and the audio detach docs, read 2026-10-03; confirm them in GET /v1/catalog before budgeting.
What does each line buy?
| Line | List rate | Units for $50 |
|---|---|---|
| Music Router tracks | $0.125 each | 400 |
| Speech-to-text minutes | $0.01 a minute | 5,000 |
| TTS characters | $0.0475 per 1,000 | 1,052,631 |
| Standalone caption jobs (up to 60 s) | $0.20 each | 250 |
| Audio detach or audio split jobs | $0.01 each | 5,000 |
| Fabric 720p audio seconds | $0.1875 a second | 266 |
How do I split a real budget?
Nobody buys only one line. A voiced, scored, captioned clip spends on several: a script's characters, a music track, a caption job and maybe a render. Work backward from the finished deliverable, price one unit of it, then multiply. The table is a ceiling check that shows which line is cheap enough to ignore (detach, split) and which dominates (lip-sync video by the second).
What is missing?
- Retakes. Every accepted generation is billed, so budget extra takes for music and narration.
- Timeline renders and video generation, which are priced per output minute or second and dwarf most audio lines.
- Anything that failed. Check the job's status before assuming a failed submit billed or did not.
Sources
Related posts
More in Pricing
- Finishing costs: captions and render as a share of a clip
Captions ($0.20) plus a one-minute render ($0.10) add $0.30 to every finished clip. That is 77% of a Grok clip's bill and 9% of a Seedance 2.5 clip's.
- Your first $10 on Sume: model, resolution and length tests
Three experiments on a $10 balance, on model, resolution and clip length, plus captions on the winner cost $8.06 and leave $1.94 for retries.
- 5 minute vertical episode: Seedance 2.5 $173.84 or avatar plus $74
A 5 minute vertical episode: $173.34 in Seedance 2.5 shots or $73.50 in avatar plus clips, plus $0.50 to join. What each route can and cannot show.
- FLUX.2 pro 1920x1080 costs $0.045 on fal: the arithmetic
fal gives $0.045 for a 1920x1080 FLUX.2 pro image and $0.03 for 1024x1024. The rounding rule behind those numbers, and how to compare it with Sume.
Written by Sume