[voice] Ask the Librarian microphone fails with No STT provider available for model whisper #144

Open
opened 2026-05-31 14:40:40 +00:00 by mik-tf · 0 comments
Owner

Clicking the microphone on the Ask the Librarian box and speaking returns a transcription error: No STT provider available for model whisper. The transcription path still routes through the old shared speech broker with a hardcoded whisper model that is not configured. It should transcribe directly against the tester's own provider key, the same way the text AI summary now does. Note that among the providers offered in the cockpit only Groq supports Whisper speech to text (OpenRouter does not transcribe audio), so when no Groq key is set the UI should show a clear message asking the user to add one rather than a generic RPC error.

Clicking the microphone on the Ask the Librarian box and speaking returns a transcription error: No STT provider available for model whisper. The transcription path still routes through the old shared speech broker with a hardcoded whisper model that is not configured. It should transcribe directly against the tester's own provider key, the same way the text AI summary now does. Note that among the providers offered in the cockpit only Groq supports Whisper speech to text (OpenRouter does not transcribe audio), so when no Groq key is set the UI should show a clear message asking the user to add one rather than a generic RPC error.
Sign in to join this conversation.
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
lhumina_code/hero_books#144
No description provided.