Telegram voice message transcription #851

Closed
opened 2026-10-27 13:38:00 +00:00 by sam · 0 comments
Owner

Hermes has STT enabled (local whisper + OpenAI whisper). Sam doesn't need full STT, but wants Telegram voice messages transcribed to text for the agent. Telegram voice messages arrive as audio attachments (voice note type, OGG/Opus format).

Scope:

  • Detect voice message type from MessageEvent
  • Transcribe using OpenAI whisper API (or local whisper via Ollama)
  • Inject transcription into message text before routing to LLM
  • Cache transcriptions to avoid re-transcribing duplicate messages

Close reason

Closed


Imported from beads archie-core-1786637498223-319-4fbc4c52 · GitHub #340 · type feature · priority P2 · created 2026-08-08 · closed 2026-10-02

Hermes has STT enabled (local whisper + OpenAI whisper). Sam doesn't need full STT, but wants Telegram voice messages transcribed to text for the agent. Telegram voice messages arrive as audio attachments (voice note type, OGG/Opus format). Scope: - Detect voice message type from MessageEvent - Transcribe using OpenAI whisper API (or local whisper via Ollama) - Inject transcription into message text before routing to LLM - Cache transcriptions to avoid re-transcribing duplicate messages ### Close reason Closed --- _Imported from beads `archie-core-1786637498223-319-4fbc4c52` · GitHub #340 · type feature · priority P2 · created 2026-08-08 · closed 2026-10-02_
sam added the mediumfeature labels 2026-10-27 13:38:00 +00:00
sam self-assigned this 2026-10-27 13:38:00 +00:00
sam closed this issue 2026-10-27 13:38:01 +00:00
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: sam/archie#851