/whatsmeow-transcribe

Service app for transcribing WhatsApp voice messages. Powered by whatsmeow and openai/whisper.

Primary LanguageGoMozilla Public License 2.0MPL-2.0

whatsmeow-transcribe

This is a small service app for transcribing (speech-to-text) WhatsApp voice messages. It is powered by whatsmeow and the openai/whisper API.

  1. Clone the repository.
  2. Run go build inside this directory.
  3. Run ./whatsmeow-transcribe --api-key sk-proj-YOUR-API-KEY-HERE to start the program.
  4. On the first run, scan the QR code. On future runs, the program will remember you (unless whatsmeow.db is deleted).

Any voice message sent to your account will be transcribed. The speech-to-text result is automatically posted to the conversation for everyone to see.

Screenshot

You can also use the API_KEY environment variable to supply the API key.
In case you are running a local text-to-speech instance, you can have --api-url point to your server.

This is a proof of concept. No support is provided.