Gemini TTS caps 200 voices per project: picking a voice on Sume
Gemini TTS stores at most 200 voices per project for a year. On Sume you list avatars and use one whose voice.status is ready, with no stored-voice quota.
Gemini's speech docs allow at most 200 stored voices per project, each with a 1-year TTL. Sume's docs state no such quota. You choose a voice by listing your avatars and calling TTS with one whose voice.status is ready.
Read 2026-10-01 from the Gemini speech page and the Sume OpenAPI description of the TTS request.
What does the 200-voice cap mean in practice?
It applies to stateful voice_ ids. Once a project holds 200, you prune before adding, and each id lapses after a year. Stateless keys sidestep the count but last 7 days and live in your client.
How do I pick a voice on Sume?
List the workspace avatars, look for voice.status equal to ready, then pass that avatar to POST /v1/tts-1.0/generate. The schema calls the avatar reference the discoverable selector.
curl https://api.sume.com/v1/avatar-1.0/avatars \
-H "Authorization: Bearer $SUME_API_KEY"Which request fields select a voice?
| Field | What it takes |
|---|---|
avatar_id | Workspace avatar id from the avatar list |
avatar_handle | Avatar handle |
voice.id | Voice UUID or library id |
voice.mode | Only id is supported |
Do I need a registry of voices in my app?
Not for quota reasons. Keep a cache of avatar ids you use often, and refresh it from the list when a call fails. If you are porting from Gemini, swap the pruning job for one list call. For the id format see Sume voice ids vs Gemini voice design.
Sources
Related posts
More in Developers
- Text to speech mulaw 8000 Hz: Gemini 3.8 and Sume TTS
Sume TTS can return pcm_mulaw or pcm_alaw at 8000 Hz in wav or raw containers. Gemini 3.8 TTS does it with audio/mulaw and audio/alaw mime types.
- Gemini voice replication: consent audio and a 10-30 s sample
Gemini API voice replication needs a 10-30 second source_audio and a consent_audio from the same speaker. Sume's TTS call takes a ready voice id instead.
- Gemini voice_key with store=false: 7 days; voice_ ids: 200, 1 year
Gemini replicated voices: stored voice_ ids keep 1 year with 200 per project; stateless voicekey_ values last 7 days. What Sume's TTS voice id accepts.
- Lost a webhook signing secret? Gemini shows it once, Sume re-reads
Gemini static webhooks return the signing secret once at creation. Sume's secret can be read again from the dashboard or GET /v1/webhooks/signing-secret.
Written by Sume