meetily

Meetily is an open-source, self-hosted AI meeting assistant built in Rust that delivers live speech-to-text transcription and local meeting summaries without sending audio or text data to the cloud.

1. For the Non-Technical Reader

Imagine having a private executive assistant sitting directly inside your laptop. They take precise, real-time notes during your meetings, organize action items, and summarize decision points, but their notebook never leaves your device. Because all processing happens on your computer, your confidential discussions, strategy sessions, and client calls stay 100% private and protected from data breaches or cloud vendor surveillance.

2. For the Technical Reader

Meetily combines a high-performance Rust architecture with local AI inference engines for low-latency audio processing:

• Transcription & Diarization: Utilizes optimized Whisper and Parakeet models for real-time live transcription running up to 4x faster, paired with local speaker diarization.
• Local Summarization: Interfaces directly with Ollama to run open-weight LLMs locally for summary generation, while retaining modular support for external providers (Claude, Groq, OpenRouter, or OpenAI-compatible endpoints).
• Platform & Portability: Cross-platform support across macOS, Windows, and Linux with full offline capability and zero cloud dependency.

3. Why It Matters

With data breach costs averaging millions and regulatory scrutiny (GDPR, HIPAA) tightening, enterprise, legal, defense, and healthcare sectors cannot afford cloud speech platforms that retain or train on sensitive audio. Meetily challenges proprietary meeting assistants by offering a self-hosted, local-first alternative that eliminates recurring per-seat SaaS costs while preserving absolute data sovereignty.

4. The "Voice AI Space Lab" Idea

The Air-Gapped Field Consult Assistant: Install Meetily on an air-gapped laptop for field medical teams or legal defense investigators. During remote client interviews or emergency patient intakes, Meetily can transcribe speech in real time and prompt a local Ollama model to instantly extract critical diagnostic symptoms or legal risk factors—completely disconnected from the internet.

Check out the open-source repository on GitHub or learn more on the Meetily Website.