Skip to main content

​
Overview

The Freya Model APIs provide direct access to our Speech-to-Text (STT) and Text-to-Speech (TTS) models. These are standalone inference endpoints — separate from the platform Call API.
ModelEndpointDescription
Speech-to-Textstt.freyavoice.aiTranscribe audio files or stream real-time audio to text
Text-to-Speechtts.freyavoice.aiConvert text to natural-sounding speech

OpenAPI: Model API endpoints

Browse the full schema for STT and TTS endpoints.

​
Authentication

All endpoints require a Bearer token issued by your workspace. See the Authentication guide for how to generate an API key. 
Authorization: Bearer YOUR_API_KEY

​
Quick Start

​
Speech-to-Text (File Upload)

curl -X POST "https://stt.freyavoice.ai/v1/audio/transcriptions" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -F "file=@audio.wav"

​
Speech-to-Text (Real-Time Streaming)

Connect via WebSocket for live transcription with partial results: 
wss://stt.freyavoice.ai/v1/audio/stream?token=YOUR_API_KEY
See the STT documentation for the full streaming wire protocol and code examples.

​
Text-to-Speech

curl -X POST "https://tts.freyavoice.ai/v1/audio/speech" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"input": "Merhaba, Freya ile konuşuyorsunuz.", "voice": "leyla", "model": "tts-1"}'
These Model APIs are independent of the Freyavoice Call API. They can be used standalone for any speech processing task, not just within Freya call workflows.