Niches · Voice cloning & custom voices
Design a new voice from a written description
No company's docs show a full fit yet. The listings below do the job with a limit, or not at all.
What counts as done
One call returns several candidate previews from a description, a chosen preview is saved as a reusable voice id in a second call, the preview audio is returned as a file or URL rather than inline encoded audio, the licence on a designed voice (including using its output to train or clone further) is stated, and the saved voice works with the vendor's main TTS models.
How agents ask for it: make a voice that sounds like a gruff old sailor
, design a narrator voice for my novel's characters
, generate a voice that isn't anyone real
, give me three voice options for this character
Does the job, with a limit 6
It does the job, but something in the docs or a user's account will get in an agent's way. The limit is named in each row.
| Listing | Why |
|---|---|
| ElevenLabs Voice Cloning and Voice Design API A Eleven Labs Inc. | POST /v1/text-to-voice/design returns several previews with generated_voice_id, saved by a second call; audio_base_64 comes inline unless stream_previews is set. The licence on designed voices sits in terms under /legal/, which robots.txt disallows. source |
| Fish Audio Voice Cloning API B Hanabi AI Inc. | POST /v1/voice-design returns several candidates and create model accepts their design signatures to save one, but candidate audio comes back as base64 WAV in the JSON body. Licence on designed voices not read. source |
| Gradium Voice Cloning and Voice Design B Gradium SAS | POST /voice-generator/generate returns several candidates, auditioned through ordinary TTS and converted into a voice used everywhere; unconverted candidates cannot be recovered by re-running. The licence on designed voices was not read. source |
| Inworld Instant Voice Cloning B Theai, Inc. dba Inworld AI | Voice Design returns up to three previews to save through the Voices API, and ad-hoc design (preview) speaks from a prompt without saving. Audio comes back base64 in audioContent, and the licence on designed voices was not read. source |
| MiniMax Voice Clone and Voice Design C Nanonoble Pte. Ltd. | Voice design takes a prompt and preview_text and creates one voice_id per call rather than several candidates, returns preview audio as hex trial_audio in the JSON, and prices previews at $30 per million characters here against $60 on the pricing page (listing notes). source |
| VoiceStudio D VoiceStudio maintainers | describe_voice maps a description to attributes without saving and design_voice saves one fixed-seed profile, so one voice rather than several candidates; output returns as a file URL in files mode. Licence depends on the engine chosen. source |
Couldn't check 2
The docs were closed to us, or silent. These carry no mark either way.
| Listing | Why |
|---|---|
| Murf Voice Cloning API C Murf Inc. | Murf's terms bar automated access to murf.ai, so its API docs were not read this run, as in the 5 November check; no first-hand account names Murf. |
| Resemble AI Voice Cloning API D Resemble AI, Inc. | No voice-design page found in the docs index read. source |
Doesn't do this job 6
Listings in the category that do something else, with the reason.
| Listing | Why |
|---|---|
| Cartesia Voice Cloning API + MCP C Cartesia AI, Inc. | No voice-design endpoint in the docs index; voices are cloned or picked from the library. source |
| Hume Octave Voice Design and Cloning + MCP C Hume AI, Inc. | Hume's docs index now lists only Audio Decisions and Expression Measurement, and the voice overview and changelog redirect there; the TTS and EVI APIs, cloning and design included, are no longer documented after the 13 November 2026 end of access. source |
| Smallest.ai Instant Voice Clone C Smallest, Inc. | No voice-design endpoint; only instant cloning from a sample. source |
| Soniox Voice Cloning BB Soniox Inc. | No voice design; instant clones from a clip only. source |
| Speechify API Voice Cloning A Speechify, Inc. | No voice design; clones come from a consented sample. source |
| Ultravox Voice Cloning E Theai, Inc. dba Inworld AI | No voice design; clones come from an uploaded sample. source |
What stops agents finishing
- preview audio returned as hex or base64 in the JSON body, filling the context (MiniMax, Hume JSON endpoint)
- a saved voice cannot be tuned afterwards (Hume)
- preview priced differently in the reference and the pricing page (MiniMax)
- Hume TTS API access ended on 13 November 2026
Around it
After: voice-cloning.clip-to-voice-id, text-to-speech.directed-performance, text-to-speech.long-form-narration.
Where this comes from
Weight 9 of 100 in Voice cloning & custom voices: 4 of 34 accounts: a novelist designing voices in one tool and cloning them into another to narrate, a Voice Design error in Gemini's studio, a licence question on training from designed-voice output, and a request for synthetic voices for audioplays
The same as data: /api/v1/use-cases/voice-cloning.design-voice-from-description.json, or the MCP tool get_use_case with id voice-cloning.design-voice-from-description. Other jobs in this category: /api/v1/niches?category=voice-cloning.
For companies
Do agents find, use and choose your tools?
An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.