Every AI note taker on the market works the same way: your meeting audio leaves your laptop, gets processed on someone else’s server, and comes back as a transcript. Meetily flips that model. It transcribes, diarizes, and summarizes your calls entirely on your own machine, and it just picked up thousands of new GitHub stars in a single week because a lot of people are done trusting the cloud with what gets said in their meetings.
What Meetily Actually Does
Meetily is an open source desktop app for Windows, macOS, and Linux that sits quietly in the background during your calls and turns them into searchable transcripts and AI generated summaries. No bot joins your Zoom. No recording notification pops up for the other people on the call. It works by capturing your microphone and system audio directly from your machine, which means it works with Zoom, Google Meet, Microsoft Teams, Discord, or literally any app that produces sound.
The transcription runs on Whisper or NVIDIA’s Parakeet model, both open source, both running entirely on your hardware. Parakeet is the newer, faster option and Meetily’s own benchmarks put it at roughly four times the speed of Whisper with lower resource use. Once the meeting ends, you point the summary step at whichever AI provider you want. Ollama running locally, or Claude, Groq, OpenRouter, and any OpenAI compatible endpoint if you’d rather send just the text summary out (not the raw audio) to a hosted model.
The project is built on Tauri, with a Rust backend and a Next.js interface, and the whole thing ships as a single installable app rather than a pile of Docker containers you have to babysit.
Why It’s Blowing Up Right Now
Here’s the thing about meeting AI tools in 2026: almost every video call platform has bolted one on, and every single one of them routes your audio through a vendor’s cloud. That’s fine for a weekly stand-up. It’s a genuine liability for a lawyer discussing a case, a doctor reviewing a patient, a founder talking through cap table numbers, or anyone in a regulated industry who can’t legally let sensitive conversations touch a third party’s servers.
The 2024 IBM Cost of a Data Breach Report put the global average cost of a breach at $4.88 million, a 10 percent jump from the year before. Every extra place your data sits, including a meeting transcript sitting on a SaaS company’s server, is one more place that number can happen to you.
Meetily’s answer landed at exactly the right moment. It hit GitHub’s trending page in early July 2026 and, according to weekly star tracking, gained close to 8,000 new stars in a single seven day stretch, on top of roughly 18,000 it already had. As of this week the repository sits above 25,000 stars and 2,600 forks. That’s an unusual growth rate for a project that’s been public for well over a year, and it lines up with a broader pattern: developers and privacy-conscious teams are actively looking for local alternatives to cloud AI tools, and word is spreading fast when they find one that actually works.
How To Install Meetily (Step by Step)
You don’t need to touch a terminal to get started on Windows or macOS. Here’s the whole process.
On Windows:
- Go to the official Releases page on GitHub.
- Download the latest
x64-setup.exefile. - Run the installer and follow the prompts. No account, no email, no sign-up screen.
- Launch Meetily from your Start menu.
On macOS:
- Download the latest
.dmgfile from the same Releases page. - Open the downloaded file.
- Drag Meetily into your Applications folder.
- Open Meetily from Applications. macOS may ask you to confirm you trust the developer the first time, since the app isn’t distributed through the App Store.
On Linux: there’s no packaged installer yet, so you’ll build it from source. Clone the repository, install Rust and Node.js if you don’t already have them, then run the project’s build script. The exact commands are in the Building from Source guide, and if you’ve ever compiled a Rust project before, it’s a fifteen minute job.
Once it’s open, grant Meetily permission to access your microphone and system audio when your OS asks. That’s the only setup most people need to start recording their first meeting.
Setting Up Free, Local AI Summaries With Ollama
Transcription works out of the box. Summaries need an AI model to actually write them, and if you want to keep everything on your machine (the whole point of using Meetily in the first place), Ollama is the natural pick.
- Download and install Ollama for your operating system.
- Open a terminal and pull a model. Something in the 7 to 8 billion parameter range, like
llama3.1orqwen2.5, gives you a good balance of summary quality and speed on a normal laptop. - In Meetily’s settings, point the AI provider to your local Ollama instance. It runs on
localhostby default, so there’s nothing to configure beyond selecting it from the dropdown. - Record a test meeting, or import an old audio file, and generate a summary to confirm the connection works.
If you’d rather not run a model locally, Meetily also lets you plug in Claude, Groq, or OpenRouter using your own API key. That sends the transcript text (not the audio) to that provider for summarization, which is a reasonable middle ground if your laptop isn’t powerful enough to run a good local model but you still don’t want to hand off raw recordings.
Free Community Edition vs. Meetily Pro
Meetily is dual-licensed in the sense that the core app is free and open source forever, and a separate paid tier exists for teams that want more.
| Feature | Community Edition (Free) | Meetily Pro |
|---|---|---|
| Local transcription (Whisper/Parakeet) | Yes | Yes, enhanced models |
| AI-generated summaries | Yes | Yes |
| Audio import and re-transcription | Yes | Yes |
| Speaker diarization | In development | Planned / rolling out |
| Auto-detect and join meetings | No | Yes |
| Calendar integration | No | Yes |
| PDF/DOCX export | No | Yes |
| Price | $0 | Roughly $10/user/month, billed annually |
Worth flagging clearly since it trips people up: Meetily’s own project description leads with “speaker diarization” as a headline feature, but as of the current release (v0.4.0, June 2026), telling speakers apart is still listed as in progress for the free edition and rolling out on Pro. If knowing exactly who said what in a transcript is your main reason for wanting this tool, check the project’s latest release notes before you install, since that’s the one feature actively being finished.
How It Compares to Otter, Fireflies, and Granola
If you’ve used Otter.ai, Fireflies, or Granola, the summaries and transcripts you’ll get from Meetily’s Community Edition are genuinely comparable, since they’re all built on similar transcription and language models under the hood. The differences that actually matter come down to three things.
No bot joins your call. Otter and Fireflies work by having a bot dial into your meeting as a visible participant. Meetily reads system audio, so nobody on the call sees a notetaker join.
Your data never has to leave your machine. That’s the entire premise, and it’s the one thing none of the cloud tools can offer no matter how good their privacy policy reads.
You’re managing the software yourself. There’s no support line, no account manager, and updates happen when you download a new release rather than automatically. For a solo founder or a small team, that trade is usually worth it. For a 200-person company that wants a help desk, the paid competitors or Meetily’s Enterprise tier make more sense.
The Honest Limitations
No mobile app exists yet. Meetily is desktop-only, so if you need to capture a meeting from your phone, this isn’t your tool today.
Linux users have to build from source. There’s no native installer, and while the process isn’t hard for anyone who’s compiled software before, it’s a real barrier for anyone who isn’t.
Speaker diarization isn’t finished in the free tier. As covered above, telling voices apart in a multi-speaker transcript is still being built out.
Local summary quality depends on your hardware. A 7 billion parameter model running on a five-year-old laptop with no dedicated GPU is going to produce noticeably weaker summaries than the same model on a machine with decent VRAM, or than a hosted model like GPT-4 or Claude. If your summaries feel thin, that’s usually the model size or your hardware, not the app itself.
It’s a young, fast-moving open source project. Expect the occasional rough edge, an update that breaks something small, and a GitHub Issues page where you’re one of the people helping shape what gets fixed next.
Bottom Line
If you sit in meetings where the content genuinely can’t leave your machine, or you’re simply tired of feeding another vendor your calendar and your conversations, Meetily is worth the fifteen minutes it takes to install. The Community Edition costs nothing, does the two things that matter most (transcribe and summarize), and keeps every recording on hardware you control. It’s not going to replace Otter or Fireflies for a 500-person sales team that needs CRM integrations and a support contract. For everyone else who just wants their meetings turned into notes without anyone else’s server in the middle, it’s one of the more genuinely useful open source releases this year.
If your meetings involve anything you’d rather not hand to a cloud AI vendor, this is worth fifteen minutes of your afternoon. Share it with the one person on your team who’s been wondering if there’s a privacy-first alternative to Otter.
Just Being Resourceful! 
Leave a comment