Resemble AI's core technology is real-time voice cloning and speech-to-speech conversion — give it a voice and it can reproduce or convert audio into that voice programmatically. That's the same underlying job Versely's voice tools do, but the cloning sits inside a studio that's already generating or editing the video the cloned voice is going to be attached to, rather than a standalone API that hands back an audio file and stops there.
What that cloning and conversion path actually does inside Versely, grounded in the real tools behind it, is below.
Clone a reusable voice from one sample
clone_voice_from_audio takes a sample recording and returns a voice_id — a reusable Cartesia voice usable in every future generate_speech or change_voice call, rather than re-recording or re-uploading a sample for each new script. list_cloned_voices browses everything already saved, so cloning happens once per voice, not once per video.
Only voices you own or have explicit consent to use can be cloned — celebrity and third-party voices require verification, and disallowed use cases are blocked outright. That consent gate applies to every clone, not an opt-in checkbox checked once at signup.
Convert an existing recording, not just generate a new one
change_voice is the speech-to-speech half of the job: it converts an existing audio clip to a different Cartesia voice while keeping the original delivery, timing and performance intact. That matters when a recording's PERFORMANCE is already right and only the voice itself needs to swap — a placeholder narration replaced with the final voice, or a speaker's identity anonymized without re-recording a single line.
Because it works on audio rather than a video file directly, changing a voice IN a video means isolating the dialogue as audio, running change_voice on it, then reattaching the result with attach_audio_to_video.
From a cloned voice to a dubbed video
dub_video runs a full dubbing pipeline — translation plus voice-cloned delivery, and optional lip-sync — asynchronously: submit the job and poll get_dub_status (or browse list_dubs) for the finished file. Two engines are available: an audio-cloning engine that works on audio or video up to 30 minutes with trimming support, and a video-only engine that lip-syncs the mouth to the new language's audio but covers a shorter runtime and fewer languages.
translate-video-to-another-language is the lighter-weight version of the same tool for when only the words need translating, not a full voice-cloned re-performance.
Where the cloned voice goes next
A voice_id from clone_voice_from_audio plugs straight into generate_speech for new narration, into change_voice for re-voicing existing audio, or into generate_lipsync to animate a face speaking it — three destinations from one cloned sample, with nothing exported between steps.
How it works
1. Provide a clean sample
A clear, representative recording is what clone_voice_from_audio needs — clarity matters more than length.
2. Name it and get a voice_id back
The returned voice_id is reusable across every future generate_speech or change_voice call.
3. Generate, convert or dub with it
Write new narration in the cloned voice, re-voice an existing recording, or feed it into a full dubbing pipeline for another language.
4. Reuse it without re-cloning
list_cloned_voices browses every voice already saved, so the next video starts from the voice_id, not a new sample.
Where this lives in Versely
Clone Your Voice for Videos
/video-editing
Change the Voice in a Video
/video-editing
Dub a Video Into Another Language
/video-editing
Translate a Video Into Another Language
/video-editing
Clone my voice from a recording
/agent
Dub a video into another language
/agent
AI Voice Cloning & Text to Speech
/tools
AI Dubbing Tool
/tools
Who this fits
- Building a consistent AI presenter voice across a whole series
- Anonymizing or standardizing a speaker's identity while keeping their exact delivery
- Dubbing existing video content into new languages with a cloned voice rather than generic TTS
- Prototyping a voice for a product without re-recording every line by hand
Frequently asked questions
How does Versely compare to Resemble AI?+
Versely covers the same underlying job — cloning a voice and using it programmatically — through clone_voice_from_audio (a reusable voice_id from one sample), change_voice (converting an existing recording to a different voice, speech-to-speech), and dub_video (a full translation-plus-cloned-voice dubbing pipeline with optional lip-sync). The difference is where it sits: all three run inside the same studio generating or editing the video, image and music the voice is going into, on one shared credit balance, rather than a standalone voice API.
Do I need consent to clone a voice?+
Yes — only voices you own or have explicit consent to use can be cloned. Third-party or celebrity voices require verification, and disallowed use cases are blocked.
Can I convert a recording's voice without retyping the script?+
Yes — change_voice keeps the original delivery and timing intact and only swaps the voice itself. If the words need to change too, that calls for a new generate_speech call instead.
Can a cloned voice dub a video into another language?+
Yes — dub_video runs translation plus a voice-cloned performance (and, on the video-only engine, lip-sync) as one asynchronous job, submitted with a project_id and checked with get_dub_status.
Other alternatives on Versely
Further reading
Try it inside Versely
The all-in-one AI studio for creators. 60+ models for video, image, voice, music and lipsync in a single app.
Reviewed August 19, 2026. Facts about Resemble AI on this page are general, publicly known positioning, not pricing or feature claims — see /alternatives for how this page set is scoped. Versely capability links above are pulled from the same live data the rest of versely.studio uses, so they move when the product does.