1. Clone from a clip​

Upload ~10–30 seconds of clear, single-speaker audio:

import os, requests

with open("my-voice.wav", "rb") as f:
    res = requests.post(
        "https://api.satryx.ai/voice/clone",
        headers={"Authorization": f"Bearer {os.environ['SATRYX_API_KEY']}"},
        files={"file": f},
        data={"name": "Chidi", "description": "My narration voice"},
    )
res.raise_for_status()
voice_id = res.json()["voice_id"]   # cloned_…

The response includes an instant preview_url so you can hear the clone immediately.

2. Speak with it​

Pass the cloned_… id straight to text-to-speech:

audio = requests.post(
    "https://api.satryx.ai/voice/tts",
    headers={"Authorization": f"Bearer {os.environ['SATRYX_API_KEY']}"},
    json={"text": "Ndewo, this na my cloned voice.", "voice_id": voice_id},
)
open("clone.wav", "wb").write(audio.content)

Dial exaggeration up for livelier delivery, or cfg_weight up to hew closer to the reference timbre.

3. Dub a video in that voice​

Cloned voices drop straight into the dubbing pipeline. Map a speaker to your clone in the voice_map when you render:

voice_map = {"SPEAKER_00": voice_id, "SPEAKER_01": "preserve"}

"preserve" keeps a speaker's own voice; your cloned_… id swaps in the clone. See the Dubbing guide for the full analyze → translate → render flow.

A note on consent​

Only clone voices you own or have explicit permission to clone. Cloning is powerful; use it responsibly.

Go deeper​

• Voice cloning
• Dubbing
• Voices & languages

Why we built it​

Most text-to-speech treats African languages as an afterthought — a flat accent bolted onto an English model, deaf to tone.

Yoruba and Igbo are tonal: the same letters mean different things depending on pitch. Nigerian Pidgin has its own rhythm. Generic TTS flattens all of it. We trained VocaBusta for real African phonology and tone, with native persona voices per language — Blessing and Sunday for Pidgin, Adunni and Tunde for Yoruba, Adaeze and Emeka for Igbo, and many more.

What you get​

• Text to speech — natural speech in 12 languages and accents, with expressiveness controls (exaggeration, cfg_weight) and streaming for low-latency playback.
• Transcription — our Vocabanga ASR model for African languages, with word-level timestamps, and Whisper as a fallback.
• Voice cloning — zero-shot: clone a voice from a ~10-second clip and synthesize with it in seconds.
• Video dubbing — analyze a video (transcript + speaker diarization), translate it, and render it re-voiced — preserving each speaker's own voice or swapping in a catalog voice.

Hello, world​

curl https://api.satryx.ai/voice/tts \
  -H "Authorization: Bearer $SATRYX_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"text": "How far? Welcome to VocaBusta.", "voice_id": "vocabusta_pcm_female"}' \
  --output hello.wav

The response is raw WAV audio; synthesis metadata comes back in the X-Vox-Metadata header.

One platform, two SDKs​

VocaBusta is part of the Satryx platform, so the same first-party SDKs that wrap chat and image generation also cover voice:

• Python — pip install satryx
• TypeScript / Node — npm install @satryx/sdk

Every endpoint is plain REST, so any language with an HTTP client works too.

Pricing & access​

Voice generation is gated behind the VocaBusta subscription — activate it in Billing, then create an API key at Account → API keys. Browsing the voice catalog (GET /voice/voices) is free and ungated, so you can audition voices before you subscribe.

Get started​

• Quickstart
• Text to speech
• API reference

We can't wait to hear what you build.