Open source · Runs on your Mac · macOS native

Your meetings, remembered.

One-click recording. Automatic transcription with speaker identification. AI-powered metadata extraction. Everything organized into Obsidian notes. All processing happens on your Mac.

Terminal

Everything you need. Nothing you don't.

Record, transcribe, and organize meetings with zero configuration with on-device AI processing.

One-Click Recording

Launch from Raycast with a keystroke. Record audio-only or capture a specific window with the native macOS picker.

Auto Transcription

WhisperX delivers accurate transcription with timestamps. Works entirely on-device using your GPU or CPU.

Speaker Identification

Pyannote diarization labels who said what. Your transcript shows Speaker 1, Speaker 2, and so on.

AI Metadata

Ollama extracts summaries, action items, decisions, key points, and tags. Structured data, not just a wall of text.

Obsidian Integration

Every meeting becomes a richly-formatted Obsidian note with frontmatter, transcript, and AI insights ready to link.

Privacy First

No accounts. No telemetry. All AI processing on your Mac. Audio never leaves your machine.

Three steps. Zero friction.

Hit a keystroke, have a conversation, come back to organized notes.

1

Record

Open Raycast. Type 'record'. Pick audio or screen mode. That's it. Recording starts instantly.

2

Process

Stop the recording and Memoant takes over. WhisperX transcribes, pyannote identifies speakers, Ollama extracts metadata.

3

Organized

A beautifully formatted Obsidian note appears with your transcript, summary, action items, and decisions. Recording saved to Google Drive.

Your data stays yours.

All AI processing runs on your Mac. WhisperX transcribes audio on your hardware. Ollama extracts metadata locally. No audio is sent to external servers. Recordings save to your Google Drive. Notes go straight to your Obsidian vault. Speaker diarization uses a one-time model download from HuggingFace. The code is open source so you can verify every line.

0
Audio uploaded
0
Accounts needed
0
Data collected
100%
Open source

Built on proven tools.

No custom models. No proprietary formats. Just best-in-class open source tools wired together.

WhisperX
Ollama
pyannote
Raycast
Obsidian
ScreenCaptureKit
Google Drive
ffmpeg

One command. Done.

macOS 12.3+ required. The installer handles everything.

1Run the installer
curl -fsSL https://raw.githubusercontent.com/kylenorthup/meeting-recorder/main/install.sh | bash
2Add your HuggingFace token (optional, for speaker labels)
# Get a token from huggingface.co/settings/tokens
# Accept terms at huggingface.co/pyannote/speaker-diarization-3.1
# Add to ~/.env: HF_TOKEN=hf_your_token_here
3Grant screen recording permission
# System Settings > Privacy & Security > Screen Recording
# Enable: Terminal and/or Raycast

Need help? Check the full README for troubleshooting and configuration options.