# Generate speech

POST /v1/speech/generations converts up to 4,096 characters to privately stored MP3 audio.

~~~bash
curl https://api.byta-ai.com/v1/speech/generations \
  -H "Authorization: Bearer $BYTA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"Byta-Audio-1.1","input":"Welcome to Byta AI.","voice":"alloy","language":"en","speed":1}'
~~~

The completed JSON response includes audio.url. Fetch it with the same Bearer key. Add ?download=1 for attachment disposition.

Enabled speech models and their USD prices per 1,000 characters are returned by GET /v1/models.