Note Taker

A minimalist, always-on-top note-taking application for macOS, built in native Swift.

Features

• Text-first design: 95% text area, minimal UI
• Syntax highlighting: Actions, connections, groups, and audience tags
• Dual-source transcription: Captures both microphone and system audio with speaker attribution
• LLM action extraction: Automatically extracts actions, commitments, and expectations from transcripts
• Always available: Global hotkey (Cmd+Shift+N) and menu bar icon
• Auto-save: Changes saved automatically after 250ms

Installation

Build from Source

cd NoteTaker

# Release build
swift build -c release

# Run
.build/release/NoteTaker

Xcode (Optional)

brew install xcodegen
cd NoteTaker
xcodegen generate
open NoteTaker.xcodeproj

Note Format

Notes are stored as markdown files in ~/Documents/Notes/ with YAML frontmatter:

---
date: 2025-12-28
group: product
audience:
  - sarah
  - team
created_at: 2025-12-28T10:00:00.000Z
updated_at: 2025-12-28T10:30:00.000Z
---

#product @sarah @team

Meeting Notes

## Actions
[] Review the proposal
[x] Send follow-up email

## Connections
Sarah -> Mike
Design -/> Implementation

Keyboard Shortcuts

Shortcut	Action
Cmd+Shift+N	Show/hide window (global)
Cmd+N	New note
Cmd+,	Preferences
Option+Up	Next note (newer)
Option+Down	Previous note (older)
Option+Left	Next note with actions
Option+Right	Previous note with actions
Double-click []	Toggle action completion
Double-click ->	Toggle connection completion

Dual-Source Transcription

Captures audio from two sources simultaneously:

• Microphone → Tagged as [You] in transcripts
• System Audio → Tagged as [Other] in transcripts (captures call participants)

Uses a hybrid approach:

• SFSpeechRecognizer for microphone
• SpeechAnalyzer for system audio (macOS 26+)

LLM Action Extraction

Automatically extracts actionable items from transcripts:

• Actions: Tasks and next steps with owner and deadline
• Commitments: Promises made during discussions
• Expectations: Things people are expecting to happen

Configure via Preferences (Cmd+,) or:

defaults write NoteTaker llm_endpoint "https://api.openai.com/v1/chat/completions"
defaults write NoteTaker llm_api_key "your-api-key"
defaults write NoteTaker llm_model "gpt-4o-mini"

Requirements

• macOS 26.0 (Tahoe) or later for dual-source transcription
• macOS 13.0 (Ventura) for single-source transcription
• Microphone permission
• Screen recording permission (for system audio)
• Speech recognition permission

Project Structure

NoteTaker/
├── App/                    # App entry point and delegate
├── Models/                 # Note, Action, Connection models
├── Services/
│   ├── ActionExtraction/   # LLM-based action extraction
│   ├── FileStorage/        # Markdown file operations
│   ├── Transcription/      # Speech recognition (dual-source)
│   └── System/             # Hotkey, permissions
├── Views/
│   ├── MainWindow/         # Main window and header
│   ├── Editor/             # NSTextView with syntax highlighting
│   ├── Preferences/        # LLM configuration
│   └── StatusBar/          # Menu bar integration
└── ViewModels/             # State management

Development

cd NoteTaker

# Debug build
swift build

# Run tests
swift test

# Release build
swift build -c release