GitHub - Polkitas/whisper-matrix-plugin: When an audio message is received, the bot downloads the audio file, converts it to a numpy array, loads the Whisper model, transcribes the audio, and sends back the result as a text message. The bot uses the Matrix Client from the mautrix library to interact with Matrix.

Run this project

Install the requerimets

Set variables envs

MATRIX_USER_ID='@transcriptorbot:example.com'
MATRIX_BASE_URL='https://matrix.server.com'
MATRIX_TOKEN='syt_xyz'

Run the bot

Screenshots: