API documentation

Voixa turns text into natural speech in ten languages, clones a voice from a ten-second sample, produces hour-long podcasts in one call, and turns audio back into text with timestamps. Everything in Studio is available over one REST API, with zero-dependency SDKs for TypeScript, Python and Go.

How it works

One base URL and one header:

https://api.voixa.vovix.io/v1

Create an API key in Studio and send it as x-api-key on every request. Then:

  1. Pick a voice. GET /voices lists the built-in voices in Vietnamese, English, Chinese, Japanese, Korean, French, German, Italian, Spanish and Portuguese, plus the voices you cloned.
  2. Send text. POST /speak returns a clip. The API waits up to about 8 seconds for the audio, so short texts usually come back ready with a signed WAV URL; longer ones come back processing and you poll.
  3. Keep it organised. Every clip lands in your library. Group clips with a project, share one with a public link, or list them by where they came from (api or studio).

Guides

Quickstart

An API key, one request, and a WAV you can play — in five minutes.

Read more

SDK

The zero-dependency TypeScript client and every method it has.

Read more

Limits

Characters per call and per day, request quotas, file sizes.

Read more

Errors

Every error code the API returns, and what to do about each.

Read more

Resources

Voices

Built-in voices in ten languages, and cloning your own from a ten-second sample.

Voice design

A voice from a description — no recording. Hear three takes, keep one. Built for dubbing.

Speak

Text in, audio out. Vietnamese has a lower-cost fast path, on by default.

Podcasts

Up to 100 000 characters produced in parallel and joined into one episode.

Clips

Your audio library: projects, filters, signed URLs and public share links.

Transcripts

Audio or video to text with timestamps, SRT and VTT — files up to 5 GB.

Account

Daily allowances, what you have used today, and supported languages.

Was this page helpful?