47
Clearminutes is the only polished, cross-platform, GPU-accelerated meeting assistant where your audio is architecturally incapable of leaving your machine. Transcription, summaries, and action items all run locally by default - not as a privacy policy you have to trust, but as a design constraint that can't be bypassed.
Local by design, not by policy - Whisper runs on your hardware (Metal/CUDA/Vulkan). No audio is ever transmitted to any server.
Built-in AI model - Ships with Gemma for summaries, no API keys or accounts required. Optionally connect your own cloud provider (Claude, OpenAI, Groq, OpenRouter) if you want.
GPU-accelerated transcription - Metal on macOS, CUDA on NVIDIA, Vulkan on AMD/Intel. Not a cloud API bottleneck.
Cross-platform desktop app - Native macOS, Windows, and Linux support via Tauri + Rust, not an Electron wrapper.
Professional audio mixing - RMS-based ducking and VAD filtering produce clean recordings and accurate transcripts, even with system audio.
No bots in your meetings - Clearminutes records locally from your device. No meeting bot joins your call, no third party intercepts your audio, no one on the call even needs to know you're recording.
Zero cloud dependency - Otter, Fireflies, and Granola all send your audio to their servers. Clearminutes processes everything on-device. If the internet drops mid-meeting, it keeps working.
No subscription required for core features - Free tier includes full transcription, summaries, and meeting history with no time limits. Pro adds speaker identification, PDF export, and TickTick integration.
One-time lifetime option - No recurring billing, no vendor lock-in on your data.
Your data stays yours - All meetings stored in local SQLite. No cloud dashboard. No training on your conversations. You can verify this because the code architecture makes it impossible to phone home audio data.
Built with