Home / Blog / Handy vs PipeVoice: Two Open-Source Dictation Tools for Windows

Handy vs PipeVoice: Two Open-Source Dictation Tools for Windows

Handy and PipeVoice are both free and open source. An honest look at where each one is the better pick on Windows.

5 min readUpdated Free · Windows

Handy and PipeVoice are both free, MIT-licensed push-to-talk dictation tools that work on Windows. Handy is a cross-platform, offline-first app. PipeVoice is Windows-only and adds optional cloud engines, per-app profiles and voice commands. Neither is better for everyone, so here is where each one wins.

Quick take: Want a simple, fully local dictation app that also runs on Mac and Linux? Handy is a strong pick. Want to mix cloud and offline engines and set different behaviour per app on Windows? Try PipeVoice.

What Handy is

From Handy's GitHub README: it is licensed MIT, built with Rust and Tauri, and runs on Windows, macOS and Linux (winget on Windows, Homebrew on macOS). Audio stays on your computer: "Your voice stays on your computer. Get transcriptions without sending audio to the cloud." You start dictation with a configurable shortcut in hold-to-record or toggle mode, and the transcript is "pasted directly into whatever app you're using". It offers Whisper model variants or Parakeet V3, uses GPU acceleration when available, filters silence with local voice activity detection, and keeps a transcript history and a custom dictionary. At the time we checked (2026-09-29) the repository showed about 32.4k stars.

What PipeVoice is

PipeVoice is a tray app for Windows 10 and 11, also MIT-licensed. You hold a hotkey, speak, and release. It has three transcription engines you can switch between at runtime: Deepgram (streaming, words appear as you speak), OpenAI Whisper, and Local Whisper (offline, about 150 MB, CPU). An optional Flow mode cleans up filler words, punctuation and casing using OpenAI, Google Gemini, OpenRouter or local Ollama. Output can be typed as real keystrokes or pasted from the clipboard.

Handy vs PipeVoice: comparison

FeaturePipeVoiceHandy
PriceFreeFree
LicenseMITMIT
PlatformsWindows 10/11 onlyWindows, macOS, Linux
OfflineYes (Local Whisper, plus Ollama for cleanup)Yes, audio stays on your computer
Cloud engines with your own keyYes (Deepgram, OpenAI Whisper)Not stated in the README summary we checked
Live streaming transcriptYes with DeepgramNot stated in the README summary we checked
Local modelsfaster-whisper (default base.en)Whisper variants or Parakeet V3
Text insertionReal keystrokes, or clipboard pastePasted into the focused app
AI text cleanupYes, Flow mode (OpenAI, Google Gemini, OpenRouter or local Ollama)Yes, post-processing with a cloud or on-device model (alpha)
Per-app profilesYes (engine, cleanup, output, Enter)Not stated in the README summary we checked
Hold or toggleBothBoth
History and dictionaryLocal history (last 50), vocabulary list, word fixesTranscript history, custom dictionary

Where Handy wins

Where PipeVoice wins

PipeVoice is also Windows-first, which is why it is built for dictating into the terminal and AI coding agents. Other Windows options are in the best free dictation software roundup.

Try PipeVoice free

Push-to-talk voice typing for Windows. Free, open source, works offline. No account.

↓ Download for Windows

free forever · open source · Windows 10 & 11

FAQ

Are Handy and PipeVoice both free and open source?

Yes. Handy's GitHub README lists an MIT license, and PipeVoice is also MIT-licensed. Both are free to use.

Does Handy work on Windows?

Yes. Handy's README lists Windows, macOS and Linux, with winget available on Windows.

Which is better, Handy or PipeVoice?

It depends. Handy is cross-platform and offline-first. PipeVoice is Windows-only but adds optional cloud engines and per-app profiles. If you use only Windows and want per-app behaviour, try PipeVoice; if you also use Mac or Linux, Handy covers more.

Do Handy and PipeVoice work offline?

Yes. Handy keeps audio on your computer, and PipeVoice can run fully offline with Local Whisper plus Ollama for cleanup.

Does PipeVoice paste or type text?

Both. By default PipeVoice types real keystrokes into the focused window, and you can switch the output mode to clipboard paste.

Sources

  1. Handy on GitHub (MIT, Rust/Tauri, platforms, offline, stars)
  2. PipeVoice README (features, engines, license)
  3. Handy docs: Post-Processing