Pindrop 🎤

The only 100% open source, truly Mac-native AI dictation app

Pindrop is a menu bar dictation app for macOS that turns your speech into text—completely offline, completely private. Built with pure Swift/SwiftUI and powered by WhisperKit for optimal Apple Silicon performance.

Download Latest Release · Documentation · Contributing · Community

---

Why Pindrop?

While other dictation apps compromise on privacy, performance, or platform fidelity, Pindrop is designed specifically for Mac users who refuse to compromise.

| Pillar | What It Means | | ------------------------------ | -------------------------------------------------------------------------- | | 🍎 Mac-Native | Pure Swift/SwiftUI—not a web wrapper. Feels like Apple built it. | | 🔒 Privacy-First | 100% local transcription. Your voice never leaves your Mac. | | ⚡ Apple Silicon Optimized | WhisperKit + Core ML = 2-3x faster than generic Whisper on M-series chips. | | 🏆 100% Open Source | No freemium tiers, no "Pro" features, no lock-in. Ever. |

---

Comparison

| Feature | Pindrop | Handy | OpenWhispr | | ------------------- | -------------------------- | --------------------- | ------------------------------ | | Platform | macOS only | Windows, macOS, Linux | Windows, macOS, Linux | | Framework | Swift/SwiftUI (native) | Tauri (Rust + Web) | Tauri (Rust + Web) | | ML Engine | WhisperKit (Apple Core ML) | Generic Whisper | Generic Whisper | | Apple Silicon | Native optimization | Emulated | Emulated | | Source Code | 100% open source | 100% open source | Freemium (paid "Lazy Edition") | | Battery Impact | Minimal (native) | Higher (web runtime) | Higher (web runtime) | | Menu Bar Design | First-class native | Web-based UI | Web-based UI |

The bottom line: If you want the best dictation experience on a Mac—maximum speed, minimal battery drain, and true native feel—Pindrop is the only choice.

---

Features

• 100% Local Transcription — Runs entirely on your Mac using OpenAI's Whisper model via WhisperKit. Your voice never leaves your computer.
• Multiple Transcription Engines — Choose between WhisperKit (Core ML optimized) and Parakeet, with streaming transcription support for real-time results.
• Global Hotkeys — Toggle mode (press to start, press to stop) or push-to-talk. Works from anywhere in macOS.
• Smart Output — Text is automatically copied to your clipboard and optionally inserted directly at your cursor.
• Notes System — Full note-taking with pinning, tagging, and AI-powered title generation. Organize and revisit your transcriptions as structured notes.
• Transcription History — All your dictations are saved locally with full search. Export to JSON, CSV, or plain text.
• Multiple Model Sizes — Choose from Tiny (fastest) to Large (most accurate) depending on your needs.
• AI Enhancement (Optional) — Clean up transcriptions using any OpenAI-compatible API—completely optional and off by default.
• Custom Dictionary — Define custom word replacements and vocabulary to improve transcription accuracy for names, jargon, and specialized terms.
• Media Controls — Automatic media pausing and system audio muting during recording so your transcription stays clean.
• Auto-Updates — Sparkle-based automatic update system keeps Pindrop up to date with zero effort.
• Beautiful macOS Design — Native SwiftUI interface that feels at home on your Mac.

---

Built With

• Swift — Apple's modern, fast, and safe programming language
• SwiftUI — Declarative UI framework for truly native Mac apps
• WhisperKit — High-performance Core ML implementation of OpenAI Whisper by Argmax, Inc.
• SwiftData — Modern data persistence framework
• Just one external dependency — WhisperKit. Everything else is Apple's first-party frameworks.

Requirements

• macOS 14.0 (Sonoma) or later
• Apple Silicon (M1/M2/M3/M4) recommended for optimal performance
• Microphone access (required for recording)
• Accessibility permission (optional, enables direct text insertion; clipboard works without it)

Installation

Pindrop releases are now signed with the project's Apple Developer identity. After the app is notarized and stapled, macOS should open it normally:

1. Download Pindrop.dmg from the releases page
2. Open the DMG and drag Pindrop to Applications
3. Launch Pindrop from Applications
4. If you downloaded a build that has not been notarized yet, macOS may still warn on first launch:
   • Right-click Pindrop → Open, or
   • Open System Settings → Privacy & Security and use "Open Anyway"
5. After notarization, Pindrop should launch without the old self-signed workaround

For maintainers: the default local release flow now exports a signed app bundle before packaging the DMG. just dmg-self-signed remains available only as a fallback when Apple signing is unavailable.

Screenshots

Coming soon: Notes & History dashboard, AI Enhancement settings, Recording indicator

Building from Source

Since this is an open-source project, you can also build it yourself. Don't worry—it's straightforward.

Step 1: Clone the Repository

git clone https://github.com/watzon/pindrop.git
cd pindrop

Step 2: Open in Xcode

open Pindrop.xcodeproj

Or simply double-click Pindrop.xcodeproj in Finder.

Step 3: Build and Run

1. In Xcode, select a scheme from the toolbar (Pindrop should be selected by default)
2. Press Cmd+R or click the Run button
3. The app will compile and launch

After the first build, Pindrop will appear in your menu bar (look for the microphone icon). The app runs exclusively in the menu bar—no dock icon.

Using the Build System (Recommended)

This project includes a justfile for common build tasks. Install just if you haven't already:

brew install just

Common commands:

just build              # Build for development (Debug)
just build-release      # Build for release
just export-app         # Export a signed app for distribution
just dmg                # Export signed app + create DMG
just test               # Run tests
just dmg-self-signed    # Fallback self-signed DMG
just clean              # Clean build artifacts
just --list             # Show all available commands

Release commands (maintainers):

just release-notes 1.9.0  # Create draft release notes file at release-notes/v1.9.0.md
just release 1.9.0  # Local manual release (tests, signed DMG, notarize/staple, appcast, tag, push tag, GitHub release)

Manual Build (Alternative)

To create a distributable build manually:

xcodebuild -scheme Pindrop -configuration Release build

The compiled app will be in DerivedData/Build/Products/Release/Pindrop.app.

Exporting a Signed App

To export a Developer ID-signed app bundle:

just export-app

Creating a DMG

To create a distributable signed DMG:

just dmg

This requires create-dmg:

brew install create-dmg

The DMG will be created in dist/Pindrop.dmg.

Creating a Release

Releases are published manually from a local machine using just + gh. Use either the one-command flow or the explicit step-by-step flow.

# One command (recommended)
just release 1.9.0

Equivalent explicit steps:

# 0. Create and edit contextual release notes
just release-notes 1.9.0

# 1. Ensure tests pass
just test

# 2. Build signed release DMG
just dmg

# 3. Generate appcast.xml for the current version
just appcast dist/Pindrop.dmg

# 4. Create and push tag
git tag -a v1.9.0 -m "Release v1.9.0"
git push origin v1.9.0

# 5. Create GitHub release with notes + attach DMG + appcast.xml
gh release create v1.9.0 dist/Pindrop.dmg appcast.xml --title "Pindrop v1.9.0" --notes-file release-notes/v1.9.0.md

First Launch

When you first open Pindrop, you'll see an onboarding flow:

1. Grant Microphone Permission — Required for recording dictations
2. Download a Model — Start with "Tiny" for the fastest experience (about 75MB)
3. Set Up Your Hotkey — Default is Option+Space for toggle mode
4. You're Ready — Press your hotkey and start dictating

Usage

Recording Modes

Toggle Mode (default: Option+Space)

• Press once to start recording (menu bar icon turns red)
• Press again to stop and transcribe
• Your transcribed text appears in your clipboard immediately

Push-to-Talk

• Hold your hotkey to record
• Release to stop and transcribe
• Configure a different hotkey in Settings → Hotkeys

Output

Transcribed text is always copied to your clipboard. If you've granted Accessibility permission,