Meetily

1. What is Meetily?

Meetily is a desktop app that records your meetings, turns speech into text as people talk, and summarizes the conversation with AI. It stands in for paid cloud note-takers (the Otter.ai type of service). The difference is that the transcription runs on your own computer, so your audio and transcripts do not have to be uploaded to anyone's server.

The "Community" edition is free and open source (MIT license). No account is needed to use it. The AI summary can run fully on your machine (through Ollama, a program that runs AI models locally), or you can use an outside service such as Claude, Groq, OpenRouter or OpenAI if you have a key for it, in which case that service bills you separately.

The app is made by Zackriya Solutions. The team also sells Meetily PRO (more advanced features, paid, see https://meetily.ai/pro/ ). This guide covers only the free Community edition.

A good fit if:

• You hold online meetings (Google Meet, Zoom, Teams...) and want minutes without putting sensitive content on a cloud service.
• You use a Windows PC or an Apple Silicon Mac.
• You are willing to set up the AI summary yourself if you want summaries to stay fully on your machine.

Not a fit if:

• You use Linux and do not compile software: there is no ready-made installer for Linux, so you must build from source.
• You use an Intel Mac: the ready-made Mac file is for Apple Silicon (aarch64) only.
• You need PRO features (PDF/DOCX export, custom summary templates...).

About the name: the project used to be called meeting-minutes and is now meetily (https://github.com/Zackriya-Solutions/meetily ). Some links in the README and on the Releases page still use the old name; GitHub redirects them to the new one, so they still open.

2. What you need

Computer:

System	Requirement
Windows	Windows 10 or later, a 64-bit CPU with AVX2 support (most CPUs from the last several years; AVX-512 is not required)
macOS	An Apple Silicon Mac (M1 or later)
Linux	No ready-made installer; build from source

The Windows build uses Vulkan to speed up speech recognition; the Mac build uses Metal and CoreML automatically. The project's documentation does not state a minimum amount of RAM or disk space; if you plan to run the AI summary on your machine, allow extra RAM and storage for that model (see Ollama's pages).

Optional, only if you want summaries on your machine: Ollama, from https://ollama.com/download . Ollama runs AI language models locally on your computer.

The latest version at the time of writing is v0.4.1, released on 12 September 2026.

No account needed. You only need an API key if you choose an outside AI service for summaries (which may cost money depending on the provider).

3. Step-by-step installation

Step 1: Download the installer

Open https://github.com/Zackriya-Solutions/meetily/releases/latest , scroll to "Assets" and download:

• Windows: meetily_0.4.1_x64-setup.exe
• Apple Silicon Mac: meetily_0.4.1_aarch64.dmg

The page also has meetily_0.4.1_x64_en-US.msi (another installer format for Windows). The .exe is enough; it is the file the README points to.

Step 2: Install on Windows

1. Double-click meetily_0.4.1_x64-setup.exe.
2. Follow the installer.
3. Open Meetily from the Start menu.

What success looks like: the Meetily window opens with its recording and transcript screen.

Step 3: Install on a Mac

1. Double-click meetily_0.4.1_aarch64.dmg.
2. Drag the Meetily icon into the Applications folder.
3. Open Meetily from Applications.

If macOS asks for microphone access, allow it; without a microphone the app cannot record your voice. If you refused by mistake, go to System Settings, Privacy & Security, and turn the microphone permission back on for Meetily.

Step 4 (optional): Install Ollama for on-device summaries

1. Download and install Ollama from https://ollama.com/download
2. Pull a model to your machine. Choose the model name from the library at https://ollama.com/library and run:

ollama pull <model-name>

Replace <model-name> with the real name of the model. Bigger models need more RAM, so pick a small one if your machine is modest.

3. Open Meetily, go to Settings, choose Ollama as the AI provider for summaries and select the model you pulled. If the app asks for an address, Ollama listens at http://localhost:11434 by default.

Skip this step if you only need recording and transcription, or plan to use an outside AI service.

Step 5 (Linux only): Build from source

Linux has no ready-made installer. The project has its own guide: https://github.com/Zackriya-Solutions/meetily/blob/main/docs/building_in_linux.md . In short, from that guide: install the basic build tools, fetch the code and run the build script.

sudo apt update
sudo apt install build-essential cmake git

You also need Node.js, Rust and pnpm as described in the official guide at https://github.com/Zackriya-Solutions/meetily/blob/main/docs/BUILDING.md , and then:

git clone https://github.com/Zackriya-Solutions/meetily
cd meetily/frontend
pnpm install --frozen-lockfile
./build-gpu.sh

This is for people with programming experience; if that is not you, use Windows or a Mac.

4. First use

Try a short recording:

1. Open Meetily.
2. Check that the microphone the app picked is your own device. Meetily records your microphone and the sound playing from the computer at the same time (for example, the other person in an online meeting).
3. Start recording (the Start button or the record button on the main screen) and say a few sentences.
4. The transcript appears gradually while you speak.
5. Stop recording. You can generate an AI summary from the meeting's detail page, if you chose an AI provider in Step 4.

A small tip: speech recognition uses a Whisper or Parakeet model running on your machine; you choose the model in the app's settings. If the app offers to download a model on first use, keep the machine online and leave enough free space. The exact size is shown in the app.

Some other features, according to the project's README:

You want to	How
Transcribe an existing audio file	Use the Import feature (currently Beta)
Re-transcribe a recorded meeting with another model or language	Use the Enhance feature
Use your company's own AI service	Pick an OpenAI-compatible endpoint (Custom OpenAI) in the settings

Note: ask participants for permission before recording a meeting. Many places have rules about recording conversations.

5. Common problems and fixes

Symptom	Cause	Fix
Your voice is not recorded	The app lacks microphone permission	Mac: turn on the microphone permission for Meetily in System Settings. Windows: open the Privacy settings, Microphone section, and allow apps to use the microphone
The app does not open or closes straight away on Windows	The CPU is too old and lacks AVX2, which the installer requires	Check your CPU (for example with the CPU-Z tool) for AVX2. If it is missing, the ready-made build cannot run on that machine. If it has AVX2, try right-clicking the Meetily icon and choosing Run as administrator, which the project's documentation suggests when the app fails to start
The .dmg will not install on an Intel Mac	The ready-made file is for Apple Silicon only	There is no prebuilt version for Intel Macs; see the source-build guide in the project's docs/BUILDING.md
The summary button shows an error when using Ollama	Ollama is not running, or no model has been pulled	Open the Ollama app, run ollama pull to fetch a model, then select the model again in Meetily
Summaries are very slow	The AI model is too large for the machine	Choose a smaller model in Ollama
A Bluetooth headset is flaky on a Mac	A problem the developers have been fixing release by release	Update to the newest version (v0.4.1 lists a fix for Bluetooth microphone reliability on macOS)

If you are still stuck, look or report at https://github.com/Zackriya-Solutions/meetily/issues

6. Uninstalling and updating

Updating

The documentation does not describe Meetily as updating itself. The way to do it: go to https://github.com/Zackriya-Solutions/meetily/releases/latest , download the new version and install it over the old one (on Windows run the new installer; on a Mac drag the new app into Applications and choose Replace). Recordings and transcripts you have saved stay on your machine.

Uninstalling

• Windows: open the installed apps list in Settings, find Meetily and uninstall.
• Mac: drag Meetily from the Applications folder to the Trash.
• If you installed Ollama only for Meetily and no longer need it, uninstall Ollama like any other app.

Recordings and transcripts are stored on your machine by the app. The documentation does not name the folder, so if you want everything wiped, check the app's storage settings before you uninstall.

7. Frequently asked questions

Does Meetily cost anything? The Community edition does not: it is free and open source (MIT). PRO is a separate paid product. If you pick an outside AI service for summaries, that service may charge you.

Does it need the internet? Recording and transcription run on your machine, so they work offline. You need a connection to download the app and models, and if you use an outside AI service for summaries.

Is my data sent anywhere? According to the README, processing happens on your machine, and models, recordings and transcripts are stored locally. The exception: if you choose an outside AI provider (Claude, Groq, OpenRouter, OpenAI or your own endpoint), the text to be summarized is sent to that provider. To keep everything on your machine, choose Ollama.

Does it support Vietnamese? The official documentation does not list the languages of each speech model. Try a short recording and look at the model and language choices in the app's settings. The Enhance feature lets you re-transcribe a meeting with another language or model.

Will it run on a weak machine? The documentation gives no minimum specification. The only stated requirement is Windows 10 or later with an AVX2 CPU, or an Apple Silicon Mac. The Ollama summary needs extra resources depending on the model; on a weak machine, use a small one.