Clone a voice from an MP4 or WebM file: Sume Voices takes video, 20 MB
Sume's Voices clone accepts wav, mp3, m4a, mp4, webm and ogg up to 20 MB. How to prepare a clean clip from a video, and what the upload check rejects.
Yes: Sume's Voices library will take a video file as the source for a clone. The upload check allows wav, mp3, m4a, mp4, webm, ogg and mpeg files of 1 byte up to 20 MB, and a clone also needs a name, a gender (male, female or nonbinary) and a language tag. The clone is then used by its id as voice.id on a text to speech request. Use your own voice, or one you have written permission to clone.
What the upload check enforces
The check lives in Sume's voice asset policy, which validates the file name extension, the content type and the size before anything is stored. A file with another extension is refused with the message "Upload a wav, mp3, m4a, mp4, webm, or ogg clip." An oversized file is refused with a size message.
| Check | Rule | If it fails |
|---|---|---|
| Extension | wav, mp3, m4a, mp4, webm, ogg, mpeg | 400 with the allowed list |
| Content type | Audio types plus video/mp4 and video/webm | 400, unsupported audio content type |
| Size | Above 0 bytes, up to 20 MB | 400 with the size range |
| Source URL | https only | Rejected if not https |
| Required fields | name, gender, language | Rejected when missing |
Prepare the clip
Cut to one person speaking, without music or room echo. If the speech lives inside a longer video that is already on Sume, audio detach returns a wav or mp3 of just its audio for $0.01 per job, and you can pass range to take a clean stretch.
Choose the language tag that matches the recording. A voice tagged for one language and asked to read another triggers Sume's mismatch warning on the first request.
Sources
Related posts
More in Models
- Czech, Hungarian, Romanian, Greek TTS API: eight codes on Sume
Sonic 3.6 lists cs, sk, hu, ro, bg, hr, el and uk. Set language for Cyrillic and Greek text or the provider reads English. Sume TTS request, codes, and price.
- Danish, Norwegian, Finnish text to speech API: da, no, fi on Sume
Sonic 3.6 lists da, no and fi. Norwegian is the code no, not nb or nn. How to send each through Sume TTS, which voice id works, and what a script costs.
- DMAD 4-step MiniMax H3: 50-step baseline vs Turbo LoRA
DMAD distills MiniMax H3 from 50 steps to 4. Why that 4 is not the Turbo LoRA's 4, what the card says, what it omits, and where hosted jobs fit.
- Edit a video with a prompt and a photo: Omni edit takes no references
Gemini Omni Flash 1.1's video edit on Sume takes a video_url and a prompt only. To put a specific person from a photo into a clip, use h3-max-recast instead.
Written by Sume