“"The issue is addressed in PR #469, however the GPU usage is not on CUDA but Vulkan for more overall compatibility." — safvanatzack, maintainer, GitHub issue #456 (https://github.com/Zackriya-Solutions/meetily/issues/456)”
You know that feeling when a meeting bot joins your call and you're suddenly not sure where that recording ends up or who can replay it later? Cloud meeting assistants like Otter.ai and Fireflies.ai transcribe your calls on their own servers, and Zackriya Solutions — the team behind Meetily — points to IBM's reported $4.4M average data-breach cost and billions of euros in GDPR fines as the reason that should worry you. You still want the transcript and the summary, you just don't want a third party holding your conversation to generate it. Meetily's answer is to run the whole pipeline — recording, transcription, summarization — on your own hardware instead.
Think of Meetily as two people listening to the same call at once: one is a stenographer who only writes when someone's actually talking, and the other quietly records everything as a clean backup track. That's a real design choice — Meetily runs your audio through a voice-detector before handing it to Whisper, which the project says cuts Whisper's workload by about 70%, because it isn't wasting cycles transcribing silence. The transcript then goes to Ollama running locally on your machine, or to Claude, Groq, or another API if you'd rather use a stronger model. Everything — audio capture, transcription, and the SQLite database that stores your notes — runs inside one Tauri app; there's no server to stand up and no Docker Compose file, because the team removed that option entirely in favor of a single local binary.
If you're an engineer or a small team that already runs Ollama locally and wants meeting notes without adding a cloud vendor to your compliance list, Meetily fits — especially on macOS or Windows, where you get a prebuilt installer. It's not there yet if you need Linux with an official binary (you're building from source), an Intel Mac (only Apple Silicon .dmg builds ship), or free-tier speaker diarization — that's Pro-only and still one of the most requested missing features nine months after the ticket opened.
Meetily is worth trying if you want a fully local meeting transcriber and can live with CPU-only transcription on Windows until you manually enable GPU support. It's an 18-month-old project that just went through a full backend rewrite, has only 12 contributors, and still has open issues around GPU detection and macOS microphone permissions — treat it as beta-quality software you test on a low-stakes meeting first, not a finished product.
Deep-dive insight, Easy and Pro modes, plus action playbooks — the full breakdown is one tap away.