voiceio vs Dragon
Dragon, from Nuance (part of Microsoft), is long-established speech recognition for Windows, with locally installed editions (Professional, Legal, Law Enforcement v16), cloud editions (Professional Anywhere, Dragon Medical One) and a mobile app. It adapts to a single speaker's voice and offers full hands-free control of the computer. Dragon for Mac was discontinued in 2018 and there is no Linux version. voiceio is free, open-source dictation for Linux that also learns your voice, by fine-tuning a Whisper model on your recordings, but it is a dictation tool, not a hands-free control system.
Which should you use?
Choose Dragon if…
- You are on Windows and need hands-free control: open apps, navigate menus, mouse grid, macros and scripting.
- You need an accessibility-certified tool (Section 508) that works with screen readers such as JAWS.
- You work in legal, law enforcement or medicine and want a domain edition and vocabulary.
- Your organization needs managed deployment, shared vocabularies and enterprise support.
Choose voiceio if…
- You are on Linux, where Dragon has no version.
- You want something free and open source, with no license to buy.
- You want voice adaptation that runs on your own hardware and that you can measure:
voiceio learn evalscores every model on clips it never trained on.
Feature comparison
| voiceio 1.5 | Dragon | |
|---|---|---|
| Platforms | Linux: Wayland and X11 (GNOME, KDE, Hyprland, sway, i3). Untested best-effort path on Windows/macOS | Windows 10/11 (desktop and cloud editions); iOS and Android (Dragon Anywhere Mobile, US and Canada). No Linux; Mac discontinued 2018 |
| Price | Free, MIT | No public price on the official pages (contact sales; subscriptions for cloud editions) |
| Where speech is decoded | On your computer. Optional text-only cloud cleanup, off by default | Local for v16 desktop editions; cloud (Azure) for Anywhere and Medical One |
| Speech models | faster-whisper small (default), medium, large-v3-turbo, distil-large-v3; experimental Parakeet and whisper.cpp server | Proprietary Nuance deep-learning engine |
| Open source | Yes (MIT) | No |
| Live typing into apps | Yes: underlined preview in the focused app through IBus, corrected in place, committed when you stop | Yes: "words appear on the computer screen, instantly" |
| Custom vocabulary | Hotwords ranked by use (about 35 per decode) plus find-and-replace corrections | Custom words, import/export, shared word lists, domain vocabularies |
| Learns from you | Yes: optional LoRA fine-tune on your own recordings, on CPU, kept only if it wins on held-out clips | Yes: adapts to the user's voice and most-used words |
| Voice commands | Dictation commands only ("new line", "scratch that", "correct that") | Full hands-free operation, mouse grid, macros, scripting |
Dragon facts are from its official site, docs or repository, read on 2026-10-10 (sources below). Prices are as listed on that date and can change. voiceio facts are from its README. Spotted something out of date? Open an issue.
Questions
Is there a Dragon NaturallySpeaking for Linux?
No. Dragon's current desktop editions require Windows 10 or 11, and no Linux client is listed (checked 2026-10-10). On Linux, voiceio, Voxtype and nerd-dictation provide dictation; Talon provides voice control on X11.
Does voiceio replace Dragon's voice commands?
No. voiceio handles dictation and a few editing commands ("new line", "scratch that", "correct that"). For hands-free control of the whole computer, look at Talon.
Does voiceio learn my voice like Dragon?
In a different way. Dragon adapts its profile as you dictate. voiceio keeps your recordings (only if you opt in), trains a LoRA fine-tune of Whisper on them on your CPU, and switches to it only when it beats the current model on held-out clips.
Is Dragon still available for Mac?
No. Nuance's support site says Dragon Professional Individual for Mac was discontinued effective 22 October 2018.
Try voiceio
Free and MIT licensed. Install it with pipx on Linux, or try the dictation in your browser first: the demo on the home page runs a small speech model inside the tab.