Documentation
The developer documentation lives at docs.vagaryvoice.cloud — one place, kept with the code. This page is the way in: what to read first, which client library fits, and where the machine-readable contract is.
From nothing to a live call
The walkthrough covers credentials, a voice call held in the browser, and a first authenticated request from your own backend.
Build against a typed client
A realtime client for the browser and mobile surface your users talk to, and a Python SDK for the server that decides what happens next.
SDKs
Typed client libraries: the browser and React realtime clients, and the Python server SDK.
Read the guideRealtime client SDK
The browser and mobile surface your end-user app uses to hold a live conversation.
Read the guidePython SDK
The pip-installable server SDK for calling platform capabilities from your backend.
Read the guideRunning it day to day
The console guides cover the areas your team works in once conversations are live.
Every endpoint, rendered from the spec
The API reference is generated from the versioned OpenAPI documents in this repository rather than written alongside them, so it lists what the contract says and nothing else. Import a spec into Postman or Insomnia, or generate a client from it.
Open the API reference- AI Voice Builder Platform API v2.3.0
/contracts/api/v1/openapi.json33 endpoints - Voice Gateway API v2.0.0
/contracts/voice-gateway/v1/openapi.json5 endpoints