Currently tts() returns raw bytes; the gateway may support format conversion depending on provider.
Task:
- Probe
POST /v1/audio/speech with response_format param (mp3/ogg/wav)
- Add
format kwarg to tts() and CLI --format
- Document which providers support which formats
Why: right now Gemini returns WAV (big); mp3 would cut size ~10x.
Currently
tts()returns raw bytes; the gateway may support format conversion depending on provider.Task:
POST /v1/audio/speechwithresponse_formatparam (mp3/ogg/wav)formatkwarg totts()and CLI--formatWhy: right now Gemini returns WAV (big); mp3 would cut size ~10x.