cjpais/Handy

A free, open source, and extensible speech-to-text application that works completely offline.

Handy is a cross-platform desktop application that provides simple, privacy-focused speech transcription. Press a shortcut, speak, and have your words appear in any text field. This happens on your own computer without sending any information to the cloud.

Handy was created to fill the gap for a truly open source, extensible speech-to-text tool. As stated on handy.computer:

• Free: Accessibility tooling belongs in everyone's hands, not behind a paywall
• Open Source: Together we can build further. Extend Handy for yourself and contribute to something bigger
• Private: Your voice stays on your computer. Get transcriptions without sending audio to the cloud
• Simple: One tool, one job. Transcribe what you say and put it into a text box

1. Press a configurable keyboard shortcut: hold it to record and release to stop, or tap it to toggle recording on and off (Hold-only and Toggle-only modes are also available)
2. Speak your words while the shortcut is active
3. Release and Handy processes your speech using Whisper
4. Get your transcribed text pasted directly into whatever app you're using

The process is entirely local:

• Silence is filtered using VAD (Voice Activity Detection) with Silero
• Transcription uses your choice of models:
  • Whisper models (Small/Medium/Turbo/Large) with GPU acceleration when available
  • Parakeet V3 - CPU-optimized model with excellent performance and automatic language detection
• Works on Windows, macOS, and Linux

1. Download the latest release from the releases page or the website
   • macOS: Also available via Homebrew cask: brew install --cask handy
   • Windows: Also available via winget: winget install cjpais.Handy
     Note: The Homebrew cask and winget package are not maintained by the Handy developers.
   • Debian/Ubuntu: Install the downloaded .deb with APT so required dependencies are installed automatically:

     sudo apt install ./Handy_*.deb

     Do not use dpkg -i unless the dependencies are already installed. If you already used it, run sudo apt --fix-broken install.
2. Install the application
3. Launch Handy and grant necessary system permissions (microphone, accessibility)
4. Configure your preferred keyboard shortcuts in Settings
5. Start transcribing!

For detailed build instructions including platform-specific requirements, see BUILD.md.

We're grateful for the support of our sponsors who help make Handy possible:

                    

Control Handy from Raycast — start/stop recording, browse transcript history, manage dictionary, switch models and languages.

Source · by @mattiacolombomc

Handy includes an advanced debug mode for development and troubleshooting. Access it by pressing:

• macOS: Cmd+Shift+D
• Windows/Linux: Ctrl+Shift+D

Handy supports command-line flags for controlling a running instance and customizing startup behavior. These work on all platforms (macOS, Windows, Linux). Largely this is a beta feature.

Remote control flags (sent to an already-running instance via the single-instance plugin):

handy --toggle-transcription    # Toggle recording on/off
handy --toggle-post-process     # Toggle recording with post-processing on/off
handy --cancel                  # Cancel the current operation

Startup flags:

handy --start-hidden            # Start without showing the main window
handy --no-tray                 # Start without the system tray icon
handy --debug                   # Enable debug mode with verbose logging
handy --help                    # Show all available flags

Flags can be combined for autostart scenarios:

handy --start-hidden --no-tray

macOS tip: When Handy is installed as an app bundle, invoke the binary directly:

/Applications/Handy.app/Contents/MacOS/Handy --toggle-transcription

This project is actively being developed and has some known issues. We believe in transparency about the current state:

Using a Bluetooth headset microphone on macOS may temporarily reduce playback quality or volume while recording because Bluetooth switches to bidirectional audio. Keep your headphones as the output device and select your Mac's built-in or an external microphone in Handy to avoid this.

Shortcuts that include the fn (Globe) key only work on Apple keyboards — your Mac's built-in keyboard or an Apple external keyboard. They will never trigger on a third-party keyboard, even while it is connected to the same Mac.

This is a hardware limitation rather than a Handy bug. fn is not part of the standard USB HID keyboard specification: Apple reports it through a vendor-specific usage that macOS honors only from Apple devices, while third-party keyboards handle their Fn key entirely in firmware and send nothing to the computer. There is no event for Handy to listen for.

If you switch between a MacBook keyboard and an external one, pick a shortcut built from standard modifiers (ctrl, option, shift, command) or a regular key instead.

Text Input Tools:

For reliable text input on Linux, install the appropriate tool for your display server:

• X11: Install xdotool for both direct typing and clipboard paste shortcuts
• Ubuntu 26.04: Has Wayland display server by default. wtype does not work, you need to install ydotool and configure systemd as described here.
• Wayland: Install wtype (preferred) or dotool for text input to work correctly
• dotool setup: Requires adding your user to the input group: sudo usermod -aG input $USER (then log out and back in)

Without these tools, Handy falls back to enigo which may have limited compatibility, especially on Wayland.

Wayland Support (Linux):

• Limited support for Wayland display server
• Requires wtype or dotool for text input to work correctly (see Linux Notes below for installation)

Other Notes:

• Runtime library dependency (libgtk-layer-shell.so.0):

  • Handy links gtk-layer-shell on Linux. If startup fails with error while loading shared libraries: libgtk-layer-shell.so.0, install the runtime package for your distro:

    Distro	Package to install	Example command
    Ubuntu/Debian	libgtk-layer-shell0	sudo apt install libgtk-layer-shell0
    Fedora/RHEL	gtk-layer-shell	sudo dnf install gtk-layer-shell
    Arch Linux	gtk-layer-shell	sudo pacman -S gtk-layer-shell

  • For building from source on Ubuntu/Debian, you may also need libgtk-layer-shell-dev.

• The recording overlay is disabled by default on Linux (Overlay Position: None) because certain compositors treat it as the active window. When the overlay is visible it can steal focus, which prevents Handy from pasting back into the application that triggered transcription. If you enable the overlay anyway, be aware that clipboard-based pasting might fail or end up in the wrong window.

• Handy disables the WebKit DMA-BUF renderer by default on Linux, but preserves explicit WEBKIT_DISABLE_DMABUF_RENDERER values. For rendering workarounds, see Troubleshooting → Linux Startup Crashes or Instability.

• If Handy fails to start reliably on Linux, see Troubleshooting → Linux Startup Crashes or Instability.

• Global keyboard shortcuts (Wayland): On Wayland, system-level shortcuts must be configured through your desktop environment or window manager. Use the CLI flags as the command for your custom shortcut.

  GNOME:

  1. Open Settings > Keyboard > Keyboard Shortcuts > Custom Shortcuts
  2. Click the + button to add a new shortcut
  3. Set the Name to Toggle Handy Transcription
  4. Set the Command to handy --toggle-transcription
  5. Click Set Shortcut and press your desired key combination (e.g., Super+O)

  KDE Plasma:

  1. Open System Settings > Shortcuts > Custom Shortcuts
  2. Click Edit > New > Global Shortcut > Command/URL
  3. Name it Toggle Handy Transcription
  4. In the Trigger tab, set your desired key combination
  5. In the Action tab, set the command to handy --toggle-transcription

  Sway / i3:

  Add to your config file (~/.config/sway/config or ~/.config/i3/config):

  bindsym $mod+o exec handy --toggle-transcription

  Hyprland:

  Add to your config file (~/.config/hypr/hyprland.conf):

  bind = $mainMod, O, exec, handy --toggle-transcription

• You can also trigger Handy externally via Unix signals or the CLI flags, which lets Wayland window managers or other hotkey daemons keep ownership of keybindings:

  Action	Trigger
  Toggle transcription	pkill -USR2 -n handy or handy --toggle-transcription
  Toggle transcription with post-processing	handy --toggle-post-process

  Example Sway config:

  bindsym $mod+o exec pkill -USR2 -n handy
  bindsym $mod+p exec handy --toggle-post-process

  pkill here simply delivers the signal—it does not terminate the process.

  Behavior change: older releases also accepted SIGUSR1 for toggling transcription with post-processing. WebKitGTK — the webview engine embedded in Handy on Linux — uses SIGUSR1 internally to coordinate JavaScript garbage collection, so listening for it caused phantom recordings and interrupted dictations every few minutes (#1660). Handy no longer listens for SIGUSR1 on Linux; the post-processing toggle is still available via handy --toggle-post-process. Remove any pkill -USR1 bindings: the signal is now delivered straight to WebKit's internal handler and can crash the app.

Overlay & Pasting Issues (Linux):

• The recording overlay window can interfere with pasting transcribed text into target applications on Linux (X11)
• Solution: Open Settings > Advanced and set "Overlay Position" to "None" to disable the overlay
• Enable "Audio Feedback" (also in Advanced) if you still want audible confirmation of recording state
• Users who upgrade from older versions or import settings from other platforms may need to manually apply this change

Handy release artifacts are signed with Tauri's updater signature format. The public key is stored in src-tauri/tauri.conf.json under plugins.updater.pubkey.

To verify a release manually, set ARTIFACT to the filename you downloaded, save the pubkey value from src-tauri/tauri.conf.json to handy.pub.b64, then decode the public key and matching .sig file from base64 and verify the artifact with minisign:

# Replace with the file you downloaded
ARTIFACT="Handy_0.8.1_amd64.AppImage"

python3 - "$ARTIFACT" <<'PY'
import base64, pathlib, sys

artifact = sys.argv[1]

pub = pathlib.Path("handy.pub.b64").read_text().strip()
pathlib.Path("handy.pub").write_bytes(base64.b64decode(pub))

sig = pathlib.Path(f"{artifact}.sig").read_text().strip()
pathlib.Path(f"{artifact}.minisig").write_bytes(base64.b64decode(sig))
PY

minisign -Vm "$ARTIFACT" \
  -p handy.pub \
  -x "$ARTIFACT.minisig"

On success, minisign prints:

Signature and comment signature verified

Do not use gpg for these .sig files.

If the transcription is correct in History but Handy inserts text you copied earlier, see issue #502. With the standard clipboard paste method, Handy restores your previous clipboard after a fixed delay. Under load, the receiving application may read the clipboard only after that restoration.

1. Open Handy's settings window and press Cmd+Shift+D (macOS) or Ctrl+Shift+D (Windows/Linux) to reveal Debug.
2. On macOS and Windows, try Reliable Paste (Beta) in Debug with a clipboard paste method selected. It uses clipboard read notifications to delay restoration instead of relying on the standard fixed delay. Test it in the application where the problem occurs; it is still experimental.
3. If Reliable Paste is disabled or unavailable, increase Paste Delay (After) in Debug and test again. This controls the wait before restoring your previous clipboard. Paste Delay (Before) controls the wait before sending the paste keystroke and addresses a different part of the operation. These delay settings apply to the standard paste path, not Reliable Paste.

If the problem persists, add your Handy version, operating system, receiving application, paste method, Reliable Paste setting, and before/after delays to the existing issue. Redact private dictated text before sharing logs.

If you're behind a proxy, firewall, or in a restricted network environment where Handy cannot download models automatically, you can manually download and install them. The URLs are publicly accessible from any browser.

1. Open Handy settings
2. Navigate to the About section
3. Copy the "App Data Directory" path shown there, or use the shortcuts:
   • macOS: Cmd+Shift+D to open debug menu
   • Windows/Linux: Ctrl+Shift+D to open debug menu

The typical paths are:

• macOS: ~/Library/Application Support/com.pais.handy/
• Windows: C:\Users\{username}\AppData\Roaming\com.pais.handy\
• Linux: ~/.config/com.pais.handy/

Inside your app data directory, create a models folder if it doesn't already exist:

# macOS/Linux
mkdir -p ~/Library/Application\ Support/com.pais.handy/models

# Windows (PowerShell)
New-Item -ItemType Directory -Force -Path "$env:APPDATA\com.pais.handy\models"

Download the models you want from below

Whisper Models (single .bin files):

• Small (487 MB): https://blob.handy.computer/ggml-small.bin
• Medium (492 MB): https://blob.handy.computer/whisper-medium-q4_1.bin
• Turbo (1600 MB): https://blob.handy.computer/ggml-large-v3-turbo.bin
• Large (1100 MB): https://blob.handy.computer/ggml-large-v3-q5_0.bin

Parakeet Unified EN 0.6B (single .gguf file, recommended):

• Q8_0 (731 MB): https://huggingface.co/handy-computer/parakeet-unified-en-0.6b-gguf/resolve/main/parakeet-unified-en-0.6b-Q8_0.gguf

For Whisper Models (.bin files):

Simply place the .bin file directly into the models directory:

{app_data_dir}/models/
├── ggml-small.bin
├── whisper-medium-q4_1.bin
├── ggml-large-v3-turbo.bin
└── ggml-large-v3-q5_0.bin

For GGUF Models (.gguf files):

Place the .gguf file directly into the models directory, exactly like the Whisper .bin files above. Handy also picks up models already present in the shared Hugging Face cache (~/.cache/huggingface/hub), so a copy downloaded by another tool works without being moved.

Important Notes:

• Do not rename the .bin or .gguf files—use the exact filenames from the download URLs
• After placing the files, restart Handy to detect the new models

1. Restart Handy
2. Open Settings → Models
3. Your manually installed models should now appear as "Downloaded"
4. Select the model you want to use and test transcription

Handy can auto-discover custom Whisper GGML models placed in the models directory. This is useful for users who want to use fine-tuned or community models not included in the default model list.

How to use:

1. Obtain a Whisper model in GGML .bin format (e.g., from Hugging Face)
2. Place the .bin file in your models directory (see paths above)
3. Restart Handy to discover the new model
4. The model will appear in the "Custom Models" section of the Models settings page

Important:

• Community models are user-provided and may not receive troubleshooting assistance
• The model must be a valid Whisper GGML format (.bin file)
• Model name is derived from the filename (e.g., my-custom-model.bin → "My Custom Model")

If Handy fails to start reliably on Linux — for example, it crashes shortly after launch, never shows its window, or reports a Wayland protocol error — try the steps below in order.

1. Install (or reinstall) gtk-layer-shell

Handy uses gtk-layer-shell for its recording overlay and links against it at runtime. A missing or broken installation is the most common cause of startup failures and can manifest as a crash or a hang well before any window is shown. Make sure the runtime package is installed for your distro:

If it is already installed and you still see startup problems, try reinstalling it (e.g. sudo pacman -S gtk-layer-shell again) in case the library files were corrupted by a partial upgrade.

2. Disable the GTK layer shell overlay (HANDY_NO_GTK_LAYER_SHELL)

If installing the library does not help, you can skip gtk-layer-shell initialization entirely as a workaround. On some compositors (notably KDE Plasma under Wayland) it has been reported to interact poorly with the recording overlay. With this variable set, the overlay falls back to a regular always-on-top window:

HANDY_NO_GTK_LAYER_SHELL=1 handy

3. Configure WebKit DMA-BUF rendering (WEBKIT_DISABLE_DMABUF_RENDERER)

Handy disables the WebKitGTK DMA-BUF renderer by default on Linux because it can cause blank windows or crashes on some GPU/driver combinations. Explicit environment-variable values are preserved, including in the Nix package. To explicitly keep the renderer disabled:

WEBKIT_DISABLE_DMABUF_RENDERER=1 handy

Some users on WebKitGTK 2.54 report a flickering or partially drawn recording overlay with the renderer disabled (#2166). If affected, try opting out of the default workaround:

WEBKIT_DISABLE_DMABUF_RENDERER=0 handy

This allows WebKit to use the renderer when supported, but can reintroduce the GPU/driver-specific crashes described above. Remove the override or use =1 if that happens.

Fully quit Handy, including its tray instance, before testing either value. Running a second instance with --toggle-transcription does not change the running process's environment.

Making a workaround permanent

Set the variable in the environment of the launcher that actually starts Handy. Shell profile exports may not apply to desktop-menu or autostart launches. If you launch Handy from a .desktop file, you can prefix the Exec= line, e.g.:

Exec=env HANDY_NO_GTK_LAYER_SHELL=1 handy

Handy's built-in autostart rewrites its generated desktop entry, so editing that entry is not a durable workaround. For autostart, use an appropriate session/service environment, or disable Handy's built-in autostart and manage your own startup entry with the required variable.

If a workaround helps you, please open an issue describing your distro, desktop environment, and session type — that information helps us narrow down the underlying bug.

If the recording overlay is empty or bordered, fully quit Handy and launch a native installation with Wayland enabled:

env -u HANDY_NO_GTK_LAYER_SHELL GDK_BACKEND=wayland handy

If this works, apply GDK_BACKEND=wayland only to Handy's launcher. This workaround does not work with the 0.9.6 AppImage, which forces X11.

On Windows, Handy asks the Vulkan loader to skip implicit layers to avoid crashes caused by overlay and capture hooks (#2049). GPU acceleration remains enabled; this does not change system-wide settings.

To opt out for GPU selection or debugging tools, fully quit Handy (including the tray icon), then run both commands in the same PowerShell window:

$env:HANDY_KEEP_VULKAN_IMPLICIT_LAYERS = "1"
& "$env:ProgramFiles\Handy\handy.exe"

Adjust the executable path if needed. This override only applies to apps launched from that PowerShell session, not the Start menu. Handy also preserves any existing VK_LOADER_LAYERS_DISABLE value.

1. Check existing issues at github.com/cjpais/Handy/issues
2. Fork the repository and create a feature branch
3. Test thoroughly on your target platform
4. Submit a pull request with clear description of changes
5. Join the discussion - reach out at contact@handy.computer

The goal is to create both a useful tool and a foundation for others to build upon—a well-patterned, simple codebase that serves the community.

• Handy CLI - The original Python command-line version
• handy.computer - Project website with demos and documentation

MIT License - see LICENSE file for details.

Handy is open-source software, but the Handy name, logo, icon, and brand assets are not open-source. Unofficial forks, rewrites, and redistributions must use their own branding and must not imply endorsement or affiliation.

• Whisper by OpenAI for the speech recognition model
• ggml for an amazing cross-platform tensor library
• Silero for great lightweight VAD
• Tauri team for the excellent Rust-based app framework
• Community contributors helping make Handy better