Skip to main content

​
Endpoints

​
Create Voice Session

POST /api/v1/voice/session
Provisions a temporary OpenAI Realtime API token for WebRTC voice connection. Response: 
{
  "token": "ephemeral-openai-token",
  "expires_at": "2026-02-18T12:00:00Z",
  "model": "gpt-4o-realtime-preview"
}
The frontend uses this token to establish a direct WebRTC connection to OpenAI’s Realtime API. Apollo Brain tools are registered as function calls within the voice session.

​
Audio Transcription

POST /audio/transcribe
Transcribes audio using OpenAI Whisper API. Request: Multipart form data with audio file. Response: 
{
  "text": "Create a high priority task for the API migration",
  "language": "en",
  "duration": 3.5
}