What problem does it solve?
Many WhatsApp voice notes arrive as audio blobs that agents cannot read, making it difficult for NanoClaw to understand and respond to voice messages. This Skill automates downloading and transcribing PTT audio so the agent receives readable, searchable text instead of opaque audio attachments.
Core Features & Use Cases
- Automatic Transcription: Downloads WhatsApp PTT audio and transcribes it with OpenAI Whisper, delivering transcripts as readable message content.
- Robust Integration: Integrates into NanoClaw's WhatsApp channel, preserving existing message handling, LID translation, and group discovery.
- Safe Fallbacks & Tests: Provides clear fallback messages when the API key is missing or transcription fails, and includes unit tests and manifest for deterministic application.
- Use Case: A user sends a WhatsApp voice note in a registered group and the agent receives and responds to the message as [Voice: <transcript>] without manual steps.
Quick Start
Enable the add-voice-transcription skill, provide your OpenAI API key when prompted, and send a WhatsApp voice note to a registered group to see it transcribed.