Why I built this: I was frustrated with the current ‘bot-in-the-middle’ era of meeting apps. Most tools (Fireflies/Otter) require a ‘bot guest’ that announces itself and kills the vibe. I also found that tools like Granola were great... but they delete your raw audio after transcribing. If the AI hallucinations happen; you have no way to verify.
What makes Loreo different?
1. Contextual AI: You can chat with a single recording or an entire series of past meetings. You can ask ‘What were the pending action items from the last three syncs?’ across your whole history. 2. Audio Ownership: Loreo keeps your audio playback and allows file uploads of existing recordings. 3. Privacy/Efficiency: Built with Tauri v2 (Rust). It’s a local ~8MB app that captures system audio directly. No ‘bot guest’ required; and your audio stays under your control. 4. Modern Engine: We use ElevenLabs Scribe v2; which is significantly more accurate for non-English languages (we are currently focused on the Japanese and Korean markets).
It’s a work in progress and I’m currently building out Team Workspaces. I’d love your feedback on the cross-meeting AI interactions and the technical implementation of the local-first capture!