Overview

A minimal, keyboard-driven desktop application for offline audio transcription using whisper.cpp.

Features

  • Private - All data stays on your machine. No network calls, no telemetry.

  • air-gap ready - Deployable on systems with no internet access

  • Keyboard-driven - Global hotkeys for recording and copying transcriptions

  • System tray - Runs in background, no main window required

  • Lightweight - Pure Rust, single binary (~10-15 MB)

How It Works

Press hotkey → start recording (tray icon turns red) Press hotkey → stop and transcribe with whisper.cpp Result appears in tray menu Press hotkey → copy to clipboard

No main window required.

Quick Start

Prerequisites

  • whisper.cpp: Build whisper.cpp and download a model (base.en recommended for ~140MB)

  • System requirements: macOS, Windows, or Linux

Installation

Build from source with Cargo, or use a pre-built binary for your platform.

First Run

The app appears in your system tray (menu bar on macOS) Right-click and select “Settings” Set paths to whisper.cpp binary and model file

Usage

Recording

Press Ctrl+Alt+R (or ⌃⌥R on macOS) to start recording Tray icon turns red while recording Press the hotkey again to stop and transcribe Notification shows transcription result

Quick Copy

Press Ctrl+Alt+C (or ⌃⌥C on macOS) to copy your most recent transcription to the clipboard.

Viewing History

Right-click the tray icon to see recent transcriptions. Click any item to copy it to clipboard.

Default Hotkeys

  • Toggle recording: Ctrl+Alt+R (or ⌃⌥R on macOS)

  • Copy last transcription: Ctrl+Alt+C (or ⌃⌥C on macOS)

See SRS for complete hotkey specifications.

Building

Requires Rust toolchain and platform-specific build tools (C compiler, audio libraries).

Requires Rust toolchain (cargo build --release). See Design (SDD) for architecture details and build requirements.

Air-Gapped Deployment

Cleanroom Whisper supports deployment on systems with no internet access. All dependencies can be vendored and transferred offline via USB or other secure methods.

For detailed air-gap deployment procedures, see the AirGap Deploy project documentation.

Privacy

Cleanroom Whisper is private by architecture:

  • Zero network code in the application

  • No analytics, telemetry, or crash reporting

  • No update checks or external API calls

  • All audio and transcriptions stay on your machine

Why Cleanroom Whisper?

Cleanroom Whisper is the only cross-platform, open-source, air-gap-ready voice transcription tool optimized for quick capture workflows in privacy-sensitive environments.

vs macOS-only tools (MacWhisper, VoiceInk, Superwhisper):

  • ✅ Cross-platform: Works on macOS, Windows, and Linux with consistent UX

  • ✅ air-gap deployment ready: Vendored dependencies, no network requirements

  • ✅ System tray workflow: Background operation with global hotkeys (no main window)

  • ✅ Minimal dependencies: 8 Rust crates vs typical 20-50+ dependencies

vs cloud-based tools (Otter.ai, Fireflies, OpenAI Whisper API):

  • ✅ 100% offline: Zero network code, guaranteed data locality

  • ✅ No subscriptions: One-time build, use forever

  • ✅ Privacy guarantee: No data leaves your machine, ever

  • ✅ Works in isolated environments: Government, healthcare, finance, research

vs file transcription tools (Vibe Transcribe, Speech Note):

  • ✅ Quick capture optimized: Global hotkey workflow for voice memos

  • ✅ Instant access: SQLite history with clipboard integration

  • ✅ Live recording: Start/stop/transcribe in seconds, not minutes

Unique positioning: The only tool designed specifically for quick voice capture in air-gapped environments where data privacy is non-negotiable.

Platform Support

Platform

Support

Notes

macOS

Full

Menu bar app

Windows

Full

System tray app

Linux (KDE, XFCE, etc.)

Full

System tray app

Linux (GNOME)

Requires extension

AppIndicator extension

Linux (Wayland)

Limited hotkeys

Global hotkeys require XWayland

License

Licensed under AGPL-3.0. Commercial licenses are available for businesses and other organizations. See LICENSE.md for details.

Documentation

This README covers installation and usage. For development and technical specifications, see the documents below.

Start Here

Document

Purpose

Principles

Core design principles (read first)

Roadmap

Project status and direction

Technical Documentation

Document

Purpose

Requirements (SRS)

Detailed functional requirements

Design (SDD)

Architecture, database schema, component design

Test Plan

Test cases and procedures

CLAUDE.md

AI assistant development guidelines

Project Planning

Document

Purpose

Roadmap

MVP implementation milestones