LMNT has shut down: what to know moving to Sume TTS
LMNT's site says it has shut down. If you move to Sume TTS 1.0, voice ids do not carry over, language must be set for non-English text, and text caps at 20000.
LMNT's website now reads "LMNT has shut down", so any integration with it needs a new provider. Sume TTS 1.0 is one option, but three things do not carry over: LMNT voice names are not Sume voice ids, non-English text needs an explicit language, and text is capped at 20000 characters per request.
LMNT's status is from its site and Sume facts from the OpenAPI reference, both read 2026-10-01. The LMNT page says nothing about migration, so the steps below are about Sume only.
What do I need to change in an LMNT integration?
| Area | What Sume's OpenAPI says |
|---|---|
| Voice id | Must be a Sume TTS voice UUID or a voi_ library id, not a voice name from another TTS ecosystem |
| Bad voice id | Rejected synchronously with 400 invalid_voice_id before a job is queued or credits are reserved |
| Language | Set language for every non-English transcript; omitted defaults to English at the provider |
| Text size | Maximum 20000 characters; spaces and punctuation count |
How do I pick a voice on Sume?
Send voice.id, or use the top-level avatar_id or avatar_handle and let Sume resolve that avatar's voice at submit time. The docs point to listing avatars with GET /v1/avatar-1.0/avatars and using one whose voice.status is ready. Copy ids verbatim; do not try to reuse an LMNT voice name. See TTS invalid voice id errors.
Why does the language field matter for a migration?
If your LMNT traffic included Korean, Japanese or other non-English text, set language (for example ko, ja, en) on every request. The docs say Sume infers ko or ja from a Hangul or kana-only transcript as a fallback, but they also say to set it explicitly. See language vs locale.
What should I test first?
Run a sample of your real scripts: check that no voice id is rejected, that long inputs stay under 20000 characters, and that non-English lines come out in the right language. Because a bad voice id fails before credits are reserved, that test costs nothing.
Sources
Related posts
More in Developers
- LTX /v1/extend limits, and chaining clips with Sume frame_images
LTX's /v1/extend is 1080p only, adds 2-20 seconds, needs 73+ input frames. Sume has no extend call here: chain clips with frame_images, then join in Timeline.
- LTX keyframe interpolation is open-source only; Sume takes last_frame
LTX's KeyframeInterpolationPipeline has no API endpoint. On Sume, frame_images accepts first_frame and last_frame on models that report support in the catalog.
- LTX retake vs Sume: fix one section of a video
LTX's POST /v1/retake regenerates one time region of a video. Sume's video edit takes a whole clip, so trim the bad section, edit it, and rejoin it.
- LTX frame counts must be 8k+1; Sume durations are whole seconds
LTX frame counts must satisfy (F-1) % 8 == 0, so 30 frames is invalid. On Sume you send whole seconds and read the allowed set from the catalog, not frames.
Written by Sume