What problem does it solve?
Many WhatsApp interactions use voice notes that agents cannot read, causing missed context and limiting conversational automation; this Skill converts audio messages into text so the BioClaw agent can understand and act on spoken input.
Core Features & Use Cases
- Automatic Transcription: Integrates OpenAI Whisper to transcribe incoming WhatsApp voice notes and convert them into plain text stored with message history.
- Database Integration & Fallbacks: Stores transcribed content in the SQLite database for agent context, with configurable fallback messages when transcription fails.
- Message-Handler Integration: Hooks into the message upsert pipeline to detect voice notes, run transcription, and persist results; includes guidance to avoid orphaned container issues on restart.
- Operational Controls: Provides a local transcription configuration file to enable/disable the feature, change providers, and customize fallback content.
- Use Case: In a registered research WhatsApp group, a user sends a short voice note asking for an analysis; the agent transcribes the note, indexes it, and replies with results based on the transcribed content.
Quick Start
Send a voice note in a registered BioClaw WhatsApp group and the agent will transcribe the audio, store the text, and respond using the transcribed content.
