Default Language
Used for speech recognition and Whisper
Backend
WebGPU is faster but requires compatible hardware
Model
onnx-community/whisper-tiny · ~39 MB
–
Meetings
–
Audio (0 chunks)
Removes recorded audio. Transcripts and notes are kept.
Open Meet runs entirely in your browser. No audio, transcripts, or notes ever leave your device. Speech recognition uses the Web Speech API. Whisper retranscription uses transformers.js with local ONNX models. Summaries use Chrome's built-in Gemini Nano.