Real-time voice identity, as an API
Swap the voice to match the face. Add live voice conversion, type-to-speak and a briefed voice agent to your product with the same engine that powers our studio.
import { createVauxClient } from '@vaux/sdk';
const vaux = createVauxClient({
// Your backend mints a short-lived token; your API key never ships to the browser.
apiKeyProvider: () => fetch('/api/vaux-token', { method: 'POST' }).then((r) => r.json()),
});
const mic = await navigator.mediaDevices.getUserMedia({ audio: true });
const session = await vaux.realtime.connect(mic, {
voice: 'leo',
emotion: 'excited',
onAudioStream: (stream) => peer.addTrack(stream.getAudioTracks()[0], stream),
});
session.set({ emotion: 'calm' }); // change live
session.say('Typed lines play instantly.');
session.sound('laugh');
session.on('transcript', (t) => console.log(t.text));How auth works
Keys stay on your server
The same pattern as leading realtime video APIs: long-lived keys server-side, short-lived tokens for clients.
Your server holds the API key
Create keys in the dashboard. They never leave your backend.
Mint a short-lived token
POST /v1/tokens with a time cap. Tokens expire fast and are single-session.
The client connects
Browser, mobile or server streams audio over WebSocket with the token.
Stream voice back
Receive speech in the chosen voice, plus transcripts and events, live.
Low latency
About half a second to first audio; live speech keeps flowing while the speaker is still talking.
Voices & emotion
Presets, clones, emotion control with delivery profiles, and real recorded reactions.
Agent endpoint
Give it a brief and remote audio; it answers in the voice you choose.
Pairs with video
Output is a standard MediaStream track, ready to merge with any face or character model.
Safe by design
Server-enforced time caps, per-key limits, consent-gated cloning, abuse reporting.
Usage you can bill
Per-session usage records and a usage API so you can pass costs to your users.
Use cases
What teams build with it
Character & face-swap apps
Pair our voice with any live face or avatar model so users look and sound like their character. One stream, both swaps.
Games & virtual worlds
Voice skins for players in voice chat, and expressive voices for in-game characters that react with laughs and gasps.
Social & streaming apps
Voice filters for live rooms, short videos and streams, alongside your camera effects.
Privacy & safety
Let people speak without revealing their real voice: support lines, dating apps, whistleblowing, community moderation.
Customer support
A consistent, clear brand voice for every agent, and a briefed agent that handles simple calls by itself.
Accessibility
Natural type-to-speak voices for people who can’t or prefer not to speak, in apps and calls.
Education & training
Role-play partners and practice conversations that answer from a lesson brief, in any voice and mood.
Content & dubbing
Narration, character lines and voiceovers generated from scripts, with emotion control.
API pricing
Start free, scale when you launch
Separate from Studio plans. Live voice is billed per minute used; text per 1,000 characters.
Sandbox
Build and test
Free
- 300 credits to start (~25 min)
- 1 live session at a time
- 20 requests / minute
- All preset voices, emotions & sounds
- Community support
Starter
Launch your first feature
$49/mo
₦79,000/mo
- 6,000 credits / month (~500 min)
- 5 concurrent sessions
- 60 requests / minute
- Voice agent endpoint
- Email support
Growth
Scale with your users
$199/mo
₦319,000/mo
- 28,000 credits / month (~2,300 min)
- 25 concurrent sessions
- 300 requests / minute
- Voice cloning API
- Usage metadata per end user
- Priority support
Scale
High-volume platforms
$799/mo
₦1,279,000/mo
- 125,000 credits / month (~10,000 min)
- 100 concurrent sessions
- 1,000 requests / minute
- Dedicated onboarding
- 99.9% uptime target
Enterprise
Custom volume & SLAs
Custom
- Volume pricing
- Custom concurrency & limits
- SLA and DPA
- Dedicated engineer
Endpoints
| POST | /v1/tokens | Mint a client token for one realtime session |
| WSS | /v1/stream | Realtime session: audio in, converted voice out, live controls |
| GET | /v1/voices | List preset, catalog and your cloned voices |
| POST | /v1/voices/clone | Create a voice clone (consent required) |
| POST | /v1/tts | One-shot text-to-speech in any voice and emotion |
| GET | /v1/sounds | List soundboard sounds |
| GET | /v1/usage | Usage and credit balance for billing your own users |
API usage draws from your API credits, separate from Studio. Volume pricing, SLAs and DPAs are available on Scale and Enterprise.