Quickstart
First call in five minutes.
Create an account, mint a test key, make a request. Test keys are free and deterministic, so you can wire up CI before you ever add credit.
| 1 · Get a key | Dashboard → API keys → Create key → Test |
| 2 · Set it | export VOCALIX_KEY=vx_test_… |
| 3 · Call it | POST /v1/tts — no SDK required, it is one HTTP request |
| 4 · Assert on it | Test keys are deterministic, so CI can diff the bytes |
curl https://api.vocalix.ai/v1/tts \ -H "Authorization: Bearer $VOCALIX_KEY" \ -H "Content-Type: application/json" \ -d '{"text":"Namaste","voice_id":"meera-hi","model":"echo-1"}' \ --output reply.wav # the resolved version comes back in a header # X-Vocalix-Model-Version: echo-1.0.0
✓ $5 credit on signup, no card✓ test keys free forever✓ errors are problem+json
Sections
Where to go next.
| API reference | Every endpoint, what authenticates it, and the headers each synthesis response carries. |
| Error codes | RFC 7807 problem+json with stable code strings. Switch on code, never on prose. |
| Deterministic testing | Same inputs, byte-identical audio, zero GPU and no wallet debit. |
| OpenAI compatibility | Already on the OpenAI SDK? Change the base URL and keep your code. |
| SDKs | First-party packages are not published yet. Here is what works today. |
| Changelog | How versions move, and the guarantees that govern it. |