Roadmap

Features on the SM-AI-MODELS roadmap. Everything on this page is planned / in progress and not yet available in the production API. The endpoints below are shown for early integration planning and may change before release.

Voice Library — GET /v1/tts/voices · 🔜 Coming Soon

List all voices available to your API key (built-in and your cloned voices), so the returned voice_id can be passed as voice in a TTS request.

Planned response:

Code
 
{
  "voices": [
    {
      "voice_id": "Yara",
      "name": "Yara",
      "type": "standard",
      "gender": "female",
      "dialects": ["ar-najdi", "ar-hijazi", "ar-qassimi", "en"],
      "status": "ready"
    },
    {
      "voice_id": "Sherine",
      "name": "Sherine",
      "type": "standard",
      "gender": "female",
      "dialects": ["ar-egyptian", "en"],
      "status": "ready"
    },
    {
      "voice_id": "Myriam",
      "name": "Myriam",
      "type": "standard",
      "gender": "female",
      "dialects": ["ar-levantine", "en"],
      "status": "ready"
    },
    {
      "voice_id": "cv_8f3a91",
      "name": "My Brand Voice",
      "type": "cloned",
      "dialects": ["ar-najdi"],
      "status": "ready"
    }
  ]
}

Planned usage flow: GET /v1/tts/voices → pick a voice_id → POST /v1/tts/audio/speech {"voice": "cv_8f3a91", ...}.

Voice Cloning — 🔜 Coming Soon

Create a custom voice from your own audio samples, scoped to your API key. Cloned voices are private to the tenant's API key and returned by GET /v1/tts/voices alongside the built-in voices.

Coming Soon — these endpoints are not live yet and are not available in the API playground.

Open items (to be finalized)

• Sample-audio requirements (minutes, format)
• Training time
• Whether cloned voices support all dialects or only the dialect of the samples
• Pricing / tier gating