Laxcorp-Research/project-raven
Open-source desktop application that captures system audio and microphone during meetings, transcribes conversations in real-time via Deepgram, and provides AI assistance using Claude or OpenAI.
Not currently ranked — collecting fresh signals.
star history
Raven is an AI-powered meeting copilot built with Electron and TypeScript that runs locally on macOS and Windows. It captures dual-stream audio (system audio plus microphone), cancels echo using WebRTC AEC3, and transcribes both sides of a conversation in real-time through the Deepgram API. During meetings, it provides context-aware AI assistance powered by Claude or OpenAI models. The application features an overlay interface with a stealth mode that remains invisible during screen sharing.
Frequently asked
- What is Laxcorp-Research/project-raven?
- Open-source desktop application that captures system audio and microphone during meetings, transcribes conversations in real-time via Deepgram, and provides AI assistance using Claude or OpenAI.
- Is project-raven open source?
- Yes — Laxcorp-Research/project-raven is open source, released under the MIT license.
- What language is project-raven written in?
- Laxcorp-Research/project-raven is primarily written in TypeScript.
- How popular is project-raven?
- Laxcorp-Research/project-raven has 407 stars on GitHub.
- Where can I find project-raven?
- Laxcorp-Research/project-raven is on GitHub at https://github.com/Laxcorp-Research/project-raven.