AI can help with meeting notes, but most solutions share one problem: your meeting content gets uploaded to the cloud.
Zoom's AI Companion, Fireflies, Otter — all great, but your meetings (potentially covering trade secrets, HR discussions, product plans) pass through someone else's servers.
Meetily takes a different path: 100% local processing, no cloud.
12,000 stars on GitHub, 1,200 forks, 490 commits. Built in Rust, supports macOS and Windows, claiming to be the "#1 self-hosted, open-source AI meeting note taker."
How it works
Transcription. Uses Parakeet or Whisper models for real-time speech-to-text. Parakeet (NVIDIA's high-performance speech model) is 4x faster than standard Whisper. All transcription happens locally — audio never leaves your machine.
Speaker diarization. Automatically identifies different speakers and segments meeting content by person.
AI summarization. After transcription, calls a local Ollama instance to generate meeting summaries, action items, and key decisions. No OpenAI, no Anthropic — all local models.
Import existing recordings. Supports importing existing meeting audio files for processing.
Tech stack
Rust backend. Core processing in Rust for performance and memory safety. Speech recognition and AI summarization are compute-intensive tasks — Rust is the right choice.
Ollama integration. Uses Ollama for local LLM summarization. You can choose different model sizes, trading quality for speed.
Local vs cloud comparison
| Dimension | Meetily (local) | Fireflies/Otter (cloud) |
|---|---|---|
| Privacy | Data never leaves device | Data uploaded to third party |
| Speed | Depends on local hardware | Cloud processing, usually faster |
| Cost | Free (open source) | Subscription |
| Model quality | Limited by local models | Cloud-scale models |
Key trade-off: Local models for summarization may not match GPT-4 or Claude quality. But you get complete privacy control.
My take
Meetily represents a growing trend: the localization of AI.
As cloud AI surges forward, so do privacy and data sovereignty concerns. Meetily's answer: bring AI capabilities back local, giving users complete control over their data.
12,000 stars shows this demand is real. Especially in enterprise environments, privacy compliance is often the first consideration when choosing AI tools.
Sources:
- GitHub - Zackriya-Solutions/meetily
- GitHub Trending Rust Weekly
- Project README and documentation