Sume audio price sheet, October 2026: TTS, STT, detach, join, render

Every Sume audio step with its rate: TTS $0.0475 per 1,000 characters, STT $0.01 a minute, detach and join $0.01, render $0.10 a minute. Plus MAI prices.

4 min readSume
All posts

Sume bills audio per job or per unit, and the rates are short enough to fit in one table. Text-to-speech is $0.0475 per 1,000 characters, speech-to-text is $0.01 per audio minute, audio detach is a flat $0.01 per job, timeline audio (join or split) is a flat $0.01, and a timeline render is $0.10 per output minute, rounded up. Confirm live rates in GET /v1/catalog, which the docs name as the source of truth.

The table

Limits are the ones in the docs and the API contract.

Sume audio steps, public rates and caps (read 2026-10-04)
StepRouteRateCap
Text to speechPOST /v1/tts-1.0/generate$0.0475 per 1,000 characters20,000 characters, 1,200 s of audio
Speech to textPOST /v1/stt-1.0/transcribe$0.01 per audio minute10 minutes per job
Audio detachPOST /v1/audio-detach$0.01 per job1,800 s source, 900 s output
Join or splitPOST /v1/timeline-1.0/audio$0.01 per job20 parts or ranges, 1,800 s output
Render with audioPOST /v1/timeline-1.0/render$0.10 per output minute1,800 s

Worked example: one 60-second short

A script of 1,000 characters costs $0.0475 to voice. Rendering it to a 60-second 1080x1920 MP4 costs $0.10. The pair is $0.1475. Add a detach and a transcription at $0.01 each if you start from an existing video, which gives $0.1675. Each step is billed on its own job, so you can price a change to one step without repricing the rest.

Against the new list prices

Microsoft published MAI-Voice-2.1 at $22 per million characters, which is $0.022 per 1,000, and the Flash tier at $15 per million, $0.015 per 1,000. MAI-Transcribe-2-Streaming is $0.54 per hour through year end, or $0.009 a minute (Microsoft AI, read 2026-10-04). On the numbers alone those are cheaper per unit than Sume's $0.0475 per 1,000 characters and $0.01 a minute.

The numbers do not include the rest of a short-video workflow. Microsoft sells a speech model. Sume's rates cover the voice, the join, the detach and the final MP4 in the same account, with GET /v1/usage listing reserved, captured and refunded rows (usage docs).

What a failed job costs

A TTS job whose audio would exceed 1,200 seconds fails with tts_duration_exceeded and nothing is captured. Estimate length before you submit with the time calculator guide.

Sources

Related posts

More in Pricing

All Pricing posts

Written by Sume