Transcription and speech façades metered in audio seconds and characters

Shipped

27 செப்டம்பர், 2026 அன்று 9:42 PM UTC

Author

Kamo

Commit

6379aa7

E4 POST /api/ai/v1/audio/transcriptions and E5 POST /api/ai/v1/audio/speech, each with its internal twin, on SP01's family adapters. The WAV-header duration walk always ends; a model that lists voices gets its first when none is named and is skipped for one it does not list (Piper ignores the voice field); Speaches pcm is asked for at 24 kHz; malformed or streaming fields are refused; a text transcript is UTF-8; an upload over 25 MB is 413 JSON.