Windows-first meeting transcription, screen recording, and Meeting Assistant. Transcribe on-device with Foundry Local or use optional Azure Speech.

Meeting transcription with Foundry Local or Azure Speech	Azure Speech batch transcription for WAV, MP3, and MP4 files	Screen recording for an entire display or selected area

npm install
npm run tauri:dev

• Build with npm run tauri:build.
• Do not use npm run dev for native features.
• Builds prepare FFmpeg 8.1.2 and bundle the Foundry Local runtime.

• Foundry Local: prepare a speech model in Settings for on-device transcription.
• Azure Speech: sign in with Azure CLI and assign the Cognitive Services User role. Supports live transcription and WAV, MP3, or MP4 files up to 250 MB and two hours.
• Retranscribe: creates a new meeting, preserving the original audio and text. Optional speaker separation and local speaker names are supported.
• On Windows 11, microphone capture follows the mute state of an available active Teams call.

Azure Speech uploads audio to the cloud. Use Foundry Local to keep transcription on-device.

Ask questions during or after recording, generate recaps and action items, and follow timestamped citations back to the recording.

1. Sign in with gh auth login using a Copilot-enabled account.
2. Open Meeting Assistant. Choose an account under Account & details if needed, then click Check status and select a model.
3. Ask a question or generate a recap. Clear session removes only assistant history, not the recording or transcript.

Powered by GitHub Copilot, not Microsoft 365 Copilot. Generation sends transcript context and conversation history to GitHub; usage charges and organization policies apply. Status checks do not send meeting content. History is saved locally.

Record a display or selected area to local H.264/H.265 MP4. Choose the output folder in Video settings. To transcribe a saved video, use its transcript icon to upload audio to Azure Speech.

• Frontend: npm run build
• Offline tests: npm run test:audio and npm run test:assistant
• Rust tests: cargo test --manifest-path src-tauri/Cargo.toml --lib from a Visual Studio 2022 x64 Developer Command Prompt.

See Architecture.md for the design and docs/rust-basic.md for the Rust implementation guide.

The bundled FFmpeg 8.1.2 essentials build is static GPLv3 software from Gyan Doshi's Windows builds: https://github.com/FFmpeg.