Scribe v2 at $0.22 an hour vs what Sume bills per audio minute

ElevenLabs lists Scribe v2 at $0.22 per hour, which is $0.22 / 60 = about $0.0037 a minute. Sume's documented speech-to-text rate is $0.01 per audio minute, which is $0.60 an hour. Sume's STT 1.0 is Scribe-class transcription, billed as a feature of the video inspect job, so the two prices are not the same product: Sume is a managed bundle with hosted media, job tracking and billing, and the ElevenLabs figure is a raw API rate.

One caution on the arithmetic. Sume's policy is provider list x1.25. The docs derive $0.010 a minute from a Scribe list of $0.008 a minute ($0.48 an hour), a different figure from the $0.22 an hour on the current ElevenLabs page. Multiplying $0.22 by 1.25 gives $0.275 an hour, not $0.60, so do not read Sume's rate as a markup on Scribe v2's current list.

The numbers

How Sume bills a transcript

You ask for a transcript by sending transcribe: true to POST /v1/video-inspect for a clip already hosted on media.sume.com. The transcript adds its per-minute rate on top of the job's Modal compute reservation, and Sume captures the actual compute, never more than the hold. A 5.5% platform fee applies to customer spend. If you omit duration_seconds, Sume reserves 1 minute for the transcript; the maximum hint is 600 seconds, and the source clip can be up to 1,800 seconds.

Optional fields are language_code (a hint such as en or ko; without it STT auto-detects) and segmentation.mode: "sentence" for gap-free sentence segments. The live rate is available from GET /v1/catalog.

Which to pick

• Bulk transcription of audio you already hold, no video workflow: the raw vendor rate is lower. Scribe v2 Medical launched Sept 11 and is billed the same as Scribe v2 per the changelog.
• Transcripts attached to clips you inspect, caption or cut in Sume: the $0.01 per minute line keeps it in one job and one invoice.
• Live use needs a streaming model. Sume transcribes files, not live streams.