Description
Whisper is accurate but processes whole files after recording, so live meetings and streams get no captions. Whisper-Flow is a framework that accepts a continuous audio stream and produces incremental transcripts immediately, turning Whisper into a real-time tool.
It is published as a Python package you can pip install and build into your own speech apps or services.
Streaming:Continuous audio chunks.
Incremental output:Text as people speak.
Whisper accuracy:Uses OpenAI Whisper.
Easy integration:A pip-installable package.
It is published as a Python package you can pip install and build into your own speech apps or services.
Features
Streaming:Continuous audio chunks.
Incremental output:Text as people speak.
Whisper accuracy:Uses OpenAI Whisper.
Easy integration:A pip-installable package.
Screenshots
Tags:transcription