100% On-Device Inference
Meeting audio is transcribed by local speech-recognition runtimes on your machine. Live summaries, complete summaries, meeting Q&A, screenshot analysis, semantic embeddings, and cross-meeting retrieval are also designed to run inside the local desktop workflow. Meeting audio, transcripts, summaries, and screenshots are not sent to an external model service for inference.
Data Stored on Your Computer
Meeting artifacts—including transcripts, summaries, screenshots, chat history, and the derived meeting knowledge index—are stored locally on your device. You can back up or delete that local data at any time.
Model Downloads
When you select a speech or language model in Settings, the application retrieves the model weight files directly over HTTPS from verified open-source distribution repositories. These download requests contain no meeting content, audio, or user identifiers.
macOS System Permissions
RecapFox requests Microphone and Screen Recording permissions solely to record user-initiated meetings and capture presentation slides. The application does not record in the background when not actively initiated by the user.
Third-Party Services & Analytics
The marketing website uses basic Firebase Analytics to gauge aggregate site traffic. The desktop application itself does not track meeting contents or personal conversations.