How to Install Meetily: Complete Desktop Setup Guide (Windows & macOS)

November 5, 2025•15 min read•Tutorial

What you'll accomplish: Install Meetily on Windows or macOS in under 15 minutes, configure audio capture for any meeting platform, select the optimal transcription engine (Whisper vs Parakeet) for your hardware, set up AI summarization with local LLMs, and record your first meeting with 100% privacy.

Want to take control of your meeting transcriptions with 100% privacy? This comprehensive guide will walk you through installing Meetily-a free, open-source AI meeting assistant that processes everything locally on your device.

Prerequisites

Windows 10/11 or macOS 10.15+
8GB+ RAM (16GB recommended)
10GB free storage space
Internet connection (for initial download only)

Why Choose Meetily?

Before we dive into installation, here's what makes Meetily special:

100% Privacy: All processing happens locally-your meeting data never leaves your device
Free & Open Source: MIT licensed, no subscriptions or per-user fees
Bot-Free: Invisible to meeting participants (no "Meetily Bot" joining calls)
Works Offline: Internet only needed for initial setup
Platform-Agnostic: Works with Zoom, Teams, Meet, Discord, any meeting software
8,100+ GitHub Stars: Trusted by privacy-conscious developers worldwide

Part 1: Installation

Option A: Install on macOS

Method 1: Download from Website (Easiest)

The simplest way to install Meetily is directly from our website.

Step 1: Visit Downloads Page

Go to meetily.ai/downloads
Select macOS from the platform selector
Choose between Community (free) or Pro version
Click "Download for macOS"

Step 2: Install Application

Open the downloaded .dmg file
Drag Meetily icon to Applications folder
Eject the installer disk image

Step 3: Launch Meetily

Open Applications folder
Double-click Meetily
If security warning appears, right-click → "Open" → Confirm

Troubleshooting macOS Security:

If macOS blocks Meetily with "App cannot be opened because the developer cannot be verified":

xattr -c /Applications/Meetily.app

Then try opening again. This removes the quarantine flag from the downloaded app.

Method 2: GitHub Releases (For More Control)

For developers who want to verify releases, check changelogs, or download specific versions:

Step 1: Download Installer

Visit Meetily GitHub Releases
Look for the macOS .dmg file in the latest release
Download and unzip if needed

Step 2: Install Application

Follow the same installation steps as Method 1 above.

Option B: Install on Windows

Method 1: Download from Website (Easiest)

The simplest way to install Meetily is directly from our website.

Step 1: Visit Downloads Page

Go to meetily.ai/downloads
Select Windows from the platform selector
Choose between Community (free) or Pro version
Click "Download for Windows"
Save the .exe installer to Downloads folder

Step 2: Prepare for Installation

Before running the installer:

Right-click the downloaded .exe file
Select "Properties"
Check "Unblock" checkbox (if present)
Click "OK"

Step 3: Run Installer

Double-click the installer file
If Windows Defender SmartScreen appears:
- Click "More info"
- Click "Run anyway"
Follow installation wizard:
- Accept license agreement
- Choose installation directory (default recommended)
- Create desktop shortcut (optional)
- Click "Install"

Step 4: Launch Meetily

Click "Finish" in installer
Meetily launches automatically
Or: Find in Start Menu → "Meetily"

Troubleshooting Windows Defender:

Windows Defender may flag new releases. This is normal for open-source software without code signing certificates. You can:

Click "More info" → "Run anyway" (safe for official releases)
Add exception in Windows Defender if needed
Verify download hash matches official release (for extra security)

Method 2: GitHub Releases (For More Control)

For developers who want to verify releases, check changelogs, or download specific versions:

Step 1: Download Installer

Visit Meetily GitHub Releases
Download the Windows .exe installer from the latest release
Save to Downloads folder

Step 2: Install Application

Follow the same installation steps as Method 1 above.

Option C: Build from Source (Advanced Users)

For developers who want to customize Meetily:

Prerequisites:

Rust toolchain (rustup)
Node.js 18+
Python 3.10+
Git

Important: Read the build guide first

Before building, read the comprehensive Building Guide for detailed instructions.

Quick Build Steps:

# Clone repository
git clone https://github.com/Zackriya-Solutions/meeting-minutes.git
cd meeting-minutes/frontend

# Install dependencies
npm install

# Development builds
npm run tauri:dev          # GPU accelerated build (recommended)
npm run tauri dev          # CPU-only build

# Production builds
npm run tauri:build        # or
npm run tauri build

For detailed troubleshooting and platform-specific instructions, see the Building Guide.

Part 2: Initial Configuration

Once Meetily launches, you'll see the setup wizard.

Step 1: Welcome & System Check

The wizard verifies:

✅ Sufficient disk space (10GB+)
✅ Memory requirements (8GB+ RAM)
✅ Audio permissions granted
✅ FFmpeg availability (installed automatically if missing)

Click "Next" when all checks pass.

Step 2: Choose Transcription Engine & Model

Meetily offers two transcription engines-a unique feature no other open-source meeting assistant has:

Option 1: Parakeet Engine (Default, Recommended)

Parakeet is selected by default and provides the best balance of speed and accuracy for most users.

Best For: Real-time transcription with excellent accuracy
Performance: Optimized for fast transcription on modern hardware
Memory: Efficient with Int8 quantization (~1-3GB RAM)
Processing: Significantly faster than Whisper with lower resource usage

Available Parakeet Models:

Lightning (Recommended): Best balance of accuracy and speed for most users
Light: Lighter model for older hardware
Compact: Most compact option with minimal resource usage

Option 2: Whisper Engine (For Maximum Accuracy)

For users who need the highest accuracy and multilingual support.

Best For: Maximum accuracy, broad language support (99+ languages)
Processing: Established, well-tested models
Access: Basic models available by default, advanced Whisper models (Medium, Large) accessible in Advanced Settings

Model	Size	RAM Usage	Speed	Best For
Tiny	~75MB	~1GB	Very Fast	Quick notes, testing
Base	~140MB	~1-2GB	Fast	Casual meetings
Small	~460MB	~2-3GB	Medium	General use
Medium	~1.5GB	~4-5GB	Slower	High accuracy needed (Advanced Settings)
Large v3	~3GB	~8-10GB	Slowest	Professional/critical meetings (Advanced Settings)

Which Should You Choose?

Use Parakeet (Default) if:

You want the best overall experience (recommended for most users)
You need real-time transcription without lag
You want lower CPU usage and battery consumption
You have any modern hardware from the last 5 years

Switch to Whisper if:

You need maximum accuracy for medical, legal, or critical meetings
You require 99+ language support
You have powerful hardware and can afford slower processing
You need specific language optimizations

Default Setup:

Meetily uses Parakeet Lightning by default, which provides excellent accuracy and the fastest transcription speed.

Model Download:

After selecting your preferred engine and model, you'll be prompted to download it
Click "Download Model" to begin the one-time download
The model downloads to your local storage (no internet needed afterward)
You can change models later in Settings if needed

Step 3: Configure Audio Sources

Microphone Selection:

Click "Microphone" dropdown
Select your input device:
- Built-in Microphone (laptops)
- External USB microphone
- Headset microphone
Test: Speak and watch the level meter

System Audio Capture:

This captures audio from your meeting platform (Zoom, Teams, etc.).

macOS:

Meetily uses BlackHole or similar virtual audio device
If not installed: Click "Install Audio Driver"
Follow prompts to install virtual audio device
Restart Meetily after installation

Windows:

Meetily uses Windows WASAPI (built-in)
Click "Enable System Audio"
Grant permissions if prompted
No additional software needed

Test Audio Capture:

Play a YouTube video or music
Check "System Audio" level meter
Verify both microphone and system audio show levels

Step 4: Storage Location

Choose where Meetily saves your transcripts and recordings.

Setting Storage Location:

The storage location is accessible and configurable directly from the Meetily UI
Access it from Settings or from the meeting details page
Click "Change" or "Browse" to select your preferred folder
Choose cloud-synced folder (Dropbox, OneDrive) for automatic backup
Or external drive for extra privacy

What Gets Stored:

Your recordings are saved in the meetily-recordings folder with the following files:

transcript.md - Full meeting transcript
summary.md - AI-generated summary (if enabled)
audio.mp3 - Audio recording (optional)

You can view and change your storage location anytime from Settings or access individual meeting files from the meeting details page.

Step 5: AI Summarization (Optional)

Meetily can generate AI summaries after each meeting. This step is completely optional-transcription works without AI summaries.

Option 1: Local LLM with Ollama (Recommended for Privacy)

Step 1: Install Ollama

Visit ollama.ai and download for your platform
Install Ollama following their instructions
Verify installation: Open terminal and run ollama --version

Step 2: Download a Model

# Recommended lightweight model (800MB)
ollama pull gemma2:2b

# Or for better quality (4.7GB)
ollama pull llama3.2:latest

Step 3: Start Ollama Server

ollama serve

Keep this terminal window open while using Meetily. The server runs at http://localhost:11434

Step 4: Configure in Meetily

Open Meetily → Settings → Model Settings
Select provider: Ollama
Enter model name (e.g., gemma2:2b or llama3.2:latest)
(Optional) Custom endpoint if not using default

Option 2: API-Based LLMs

If you prefer cloud LLMs, Meetily supports:

Anthropic Claude: Settings → Model Settings → Provider: Claude → Enter API key
OpenAI GPT: Settings → Model Settings → Provider: OpenAI → Enter API key
Groq: Settings → Model Settings → Provider: Groq → Enter API key (free tier available)
OpenRouter: Settings → Model Settings → Provider: OpenRouter → Enter API key

Note: Using API-based LLMs means your meeting transcripts are sent to third-party servers for summarization.

Option 3: Skip for Now

You can enable AI summaries later in Settings → Model Settings
Transcription works perfectly without AI summaries
Manual review of transcripts works fine initially

Click "Finish" to complete setup!

Part 3: Recording Your First Meeting

Before the Meeting

1. Launch Meetily

Open from Applications (Mac) or Start Menu (Windows)
Ensure it's running (icon in menu bar/system tray)

2. Join Your Meeting

Open Zoom, Teams, Google Meet, Discord, etc.
Join meeting as normal
No special setup needed per platform

3. Start Recording

Community Edition:

In Meetily, click "Record" button to start recording

Pro Edition:

Automatic Meeting Detection: Pro users benefit from automatic meeting detection
Meetily automatically detects when you join a meeting and can start recording
Global Shortcuts: Use Option+X (Mac) / Alt+X (Win) to start/stop recording from anywhere without UI interaction
Floating Recording Panel: Access recording controls via a convenient floating panel
You can also manually start/stop recordings as needed
Configure auto-detection settings in Settings → Meeting Detection

During the Meeting

Real-Time Transcription Display:

Recording in progress...

Good morning everyone, let's start with our product updates.

Sure, I can begin. We've made great progress on the authentication feature...

Controls:

Pause: Temporarily pause recording (excludes segments from transcript)
Bookmark: Mark important moments (adds timestamp marker)
Stop: End recording (triggers AI summary)

Performance Monitoring:

CPU usage indicator
Memory usage
Transcription lag (should be <2 seconds)

After the Meeting

1. Stop Recording

Click "Stop" button in Meetily
Pro users can use global shortcut: Option+X (Mac) / Alt+X (Win)

2. AI Summary Generation (if enabled)

Processing summary...
- Analyzing transcript
- Extracting key points
- Identifying action items
- Detecting decisions

Summary complete! (takes 10-60 seconds)

3. Review & Export

Community Edition:

View transcript in Meetily
Export to Markdown
Copy to clipboard for pasting elsewhere

Pro Edition (Additional Features):

Audio Upload (Coming Soon): Upload existing audio files for transcription
Templates: Use pre-built templates for different meeting types
Chat: Interactive chat with your transcripts using AI
Integrations: Connect with your favorite tools
Cloud Sync: Sync transcripts across devices
Sharable Links: Share transcripts with team members across organizations

Part 4: Platform-Specific Configuration

Google Meet

No special setup needed!

System audio capture works automatically
Just click "Record" in Meetily before/during meeting

Tips:

Use Chrome or Edge (better audio routing)
Ensure "Use original sound" is enabled in Meet settings

Zoom

Configuration:

Open Zoom → Settings → Audio
Enable "Original Sound for Musicians"
Disable "Suppress background noise" (optional, for better accuracy)

Recording:

Start Meetily recording BEFORE joining Zoom call
Or start after joining (may miss first few seconds)

Microsoft Teams

Configuration:

Open Teams → Settings → Devices
Test your microphone and speakers
Disable "Noise suppression" for better transcription accuracy

Note: Teams' noise suppression can filter out background speakers. Disable for meetings where you need to capture all participants.

Discord

Works perfectly with system audio capture!

No special configuration needed
Captures all voice channels
Real-time transcription during gaming sessions or community calls

Slack Huddles

Configuration:

Join Slack Huddle
Start Meetily recording
System audio captures Huddle audio automatically

Part 5: Advanced Configuration

GPU Acceleration (Automatic)

Meetily automatically detects and uses GPU acceleration if available on your system. No manual configuration needed!

How It Works:

GPU acceleration is detected and enabled automatically at application startup
On macOS: Uses Metal acceleration (built into Apple Silicon M1/M2/M3)
On Windows: Uses compatible NVIDIA CUDA or AMD ROCm if available
Fallback to CPU if no compatible GPU is found

Supported GPUs:

Apple Silicon: M1, M2, M3, M4 (automatic Metal acceleration)
NVIDIA: GTX 1060 or newer (requires CUDA)
AMD: RX 580 or newer (requires ROCm on Linux)

Note: GPU acceleration provides 3-10x faster transcription compared to CPU-only processing. If you experience issues, the app will automatically fall back to CPU processing.

Global Shortcuts (Pro Only)

Pro users have access to global shortcuts that work without UI interaction:

Available Shortcuts:

Start/Stop Recording: Option+X (Mac) / Alt+X (Win)
Works from anywhere on your system, even when Meetily is minimized
No need to switch windows or interact with the UI
Perfect for seamless recording workflow

Configure and customize shortcuts in Settings → Shortcuts (Pro only)

Application Settings

Notification Preferences:

Settings → Preferences → Enable/disable recording notifications
Control which events trigger notifications
Respect Do Not Disturb mode

Storage Locations:

Settings → Preferences → View storage paths
Database location: Where meeting metadata is stored
Models directory: Where transcription models are cached
Recordings folder: Where audio and transcripts are saved

Analytics:

Settings → Preferences → Analytics consent toggle
Opt in/out of usage analytics
Helps improve the app without compromising privacy

Part 6: Troubleshooting Common Issues

Issue 1: "No Audio Detected"

Symptoms: Recording starts but no transcription appears

Solutions:

Check Audio Permissions:
- macOS: System Preferences → Security → Microphone → Enable Meetily
- Windows: Settings → Privacy → Microphone → Allow Meetily
Test Audio Sources:
- Settings → Audio → Test Microphone
- Speak and verify level meter moves
Restart Audio Devices:
- Unplug and replug headset/microphone
- Restart Meetily

Issue 2: "Transcription Too Slow"

Symptoms: Real-time transcription lags 10+ seconds behind speech

Solutions:

Use Smaller Model:
- Switch from Large → Medium or Small
- Settings → Models → Select smaller model
Enable GPU:
- Settings → Performance → Enable GPU
- Requires compatible graphics card
Close Other Apps:
- Free up RAM and CPU
- Close Chrome tabs, video editors, etc.

Issue 3: "Model Download Fails"

Symptoms: Whisper model download stuck or fails

Solutions:

Check Internet Connection:
- Ensure stable internet
- Large v3 is 3GB download (takes time)
Manual Download:
- Download from Hugging Face
- Settings → Models → "Import Model" → Select file
Clear Cache:
- Settings → Advanced → "Clear Model Cache"
- Retry download

Issue 4: "Meetily Won't Launch"

Symptoms: Double-clicking icon does nothing

macOS Solutions:

# Check if app is blocked
xattr -d com.apple.quarantine /Applications/Meetily.app

# Check for errors
/Applications/Meetily.app/Contents/MacOS/Meetily

Windows Solutions:

Run as Administrator: Right-click Meetily → "Run as administrator"
Check Event Viewer: Windows → Event Viewer → Application logs (Look for Meetily errors)
Reinstall: Uninstall via Control Panel, download fresh installer, install again

Issue 5: "Poor Transcription Accuracy"

Symptoms: Many words transcribed incorrectly

Solutions:

Upgrade Whisper Model:
- Base → Small or Medium
- Accuracy improves significantly
Improve Audio Quality:
- Use external microphone (not laptop built-in)
- Reduce background noise
- Move closer to microphone
Disable Noise Suppression:
- In Zoom/Teams settings
- Noise suppression can confuse Whisper

Part 7: Next Steps

Optimize Your Workflow

1. Set Up Ollama for Local AI (If Not Done Already)

Complete privacy: AI summaries without cloud APIs
Install from ollama.ai
Pull recommended model: ollama pull gemma2:2b
Start server: ollama serve
Configure in Meetily → Settings → Model Settings

2. Create Meeting Templates

Settings → Templates → "New Template"
Pre-fill meeting types (standup, client call, etc.)
Automatic file naming and folder organization

3. Export Workflows

Notion: Export Markdown → Import to Notion
Obsidian: Save to Obsidian vault folder
Google Docs: Copy transcript, paste directly

Join the Community

GitHub: Star the repo (8,100+ stars!)Discord: Join community

Upgrade to Teams (Optional)

For 2-100 users, Meetily Organizations tier offers:

Central server for team transcript sharing
Admin dashboard for user management
Managed deployment by Zackriya team
Extended support

For 100+ users, Meetily Enterprise adds:

Server-side AI model hosting (larger models)
Custom infrastructure (AWS/Azure/on-premise)
GDPR/HIPAA managed compliance
24/7 dedicated support

Request Demo →

Conclusion

Congratulations! You now have Meetily installed and configured. You're ready to:

✅ Record meetings with complete privacy
✅ Get accurate AI transcriptions
✅ Generate automatic summaries
✅ Export to your favorite tools
✅ All without cloud subscriptions or data uploads

Get started:

Join your next meeting
Click "Record" in Meetily
Review transcript and summary after
Enjoy privacy-first meeting intelligence!

About the Author

Written by the Meetily team. We built Meetily to provide professionals with privacy-first meeting intelligence that doesn't compromise on features or usability. Our mission is to make AI meeting transcription accessible to everyone while keeping your data completely private.

GitHub Discord Community Product Hunt

Frequently Asked Questions

How much disk space do I need?

For Meetily, you'll need approximately 10GB for initial setup (models: 1-3GB one-time download). Per hour of meeting recording: Audio file (MP3): ~50-100MB, Transcript (Markdown): ~50-200KB. For 10 meetings/week, budget ~5GB/month for audio. Tip: Enable 'Delete audio after transcription' in settings to save space while keeping transcripts.

Can I use Meetily on multiple computers?

Yes! For the free individual tier, install on multiple devices with transcripts stored locally on each device. Use cloud storage (Dropbox/OneDrive) to sync transcripts between devices. For Organizations tier, a central server syncs across all team devices, making transcripts accessible from any authenticated device.

Does Meetily work on Linux?

Not yet as a native app. Linux support is actively being developed and will be available in a future release. Currently supported platforms: Windows (native installer) and macOS (native installer or Homebrew). Linux users can use Docker for now.

What languages does Meetily support?

Whisper supports 99 languages including English, Spanish, French, German, Italian, Portuguese, Dutch, Polish, Russian, Japanese, Chinese, Korean, Arabic, Hindi, and many more! Accuracy is best for English, with strong performance for other major languages.

Is my meeting data really private?

100% yes for the individual tier. All processing happens locally on your device, transcripts are saved to your chosen folder, no data is sent to the internet (except if you use API-based AI), and the open-source code can be audited. You have complete control over your data.

Which transcription engine should I choose: Whisper or Parakeet?

Choose Whisper if you need maximum accuracy (medical, legal, critical meetings), require 99+ language support, or have powerful hardware. Choose Parakeet if you have older/slower hardware (8GB RAM or less), need real-time transcription without lag, or want lower CPU usage. For 8GB RAM: Parakeet recommended. For 16GB+ RAM: Whisper Medium or Large for best accuracy.

How do I troubleshoot 'No Audio Detected' errors?

Check audio permissions in System Preferences (macOS) or Settings (Windows). Test audio sources in Meetily Settings → Audio → Test Microphone. Restart audio devices by unplugging and replugging headset/microphone, then restart Meetily. Verify that Meetily has microphone access enabled in your operating system.

Why is transcription too slow on my computer?

Use a smaller Whisper model (switch from Large → Medium or Small in Settings). Enable GPU acceleration if you have a compatible graphics card (Settings → Performance → Enable GPU). Close other applications to free up RAM and CPU. Consider switching to the Parakeet engine which is optimized for speed.

Can I use custom AI models with Meetily?

Yes! Meetily supports multiple transcription models including Whisper variants and custom models via Ollama or llama.cpp. You can fine-tune models for domain-specific terminology (medical, legal, technical) and use custom LLMs for meeting summarization. This flexibility is only possible with open-source, self-hosted architecture.

How does Meetily work with Google Meet, Zoom, and Teams?

Meetily works with all meeting platforms through system audio capture. For Google Meet: No special setup needed. For Zoom: Enable 'Original Sound for Musicians' in Audio settings. For Microsoft Teams: Disable 'Noise suppression' for better accuracy. For Discord: No configuration needed. Just start recording in Meetily before or during your meeting.

Meetily vs Otter.ai: Privacy Comparison

Compare Meetily's self-hosted approach vs Otter.ai's cloud platform.

HIPAA Compliance Guide

Set up HIPAA-compliant meeting transcription for healthcare.