Compose a draft email to my studio professor asking to move my desk crit to Thursday.
Two modes
Hold-to-Talk for quick bursts, Hands-Free toggle for long dictation.
Axon Voice sits in your tray, listens while you hold a key, and drops clean text into whatever app you're already in. No account, no API key, no cloud.
Free forever · runs offline · your voice never touches a server
This is the actual interface — three tabs, that's the whole app. Switch tabs on the left; hover or tap any control and the panel on the right tells you what it does.
Your dictations — cleared every time you open Axon, kept only for this session.
Compose a draft email to my studio professor asking to move my desk crit to Thursday.
Write me a prompt for an image model: a brutalist concrete pavilion at golden hour, light fog.
Loft the selected curves in Grasshopper, then bake the brep to layer two.
Summarize these three readings into five bullet points I can drop into my thesis notes.
Say the keyword, get the expansion. Applied automatically before pasting.
Everything Axon does, in one screen. No menus to dig through.
Click “Change”, press any combo and it becomes your hotkey instantly. It's captured globally while Axon runs, so it won't trigger other apps' shortcuts.
Now using auto-selected model.
Every transcript from this session, newest first, each with a one-click copy. It's session-only — quit Axon and it's wiped. Nothing is written to disk or sent anywhere.
↑ hover a control in the window
No bloat, no menus you'll never open. Dictation that behaves the same in every app.
A local Whisper model does the transcription on your own machine. After the one-time model download, you can pull the plug on the internet and Axon keeps working. Audio and text never leave your computer.
Hold-to-Talk for quick bursts, Hands-Free toggle for long dictation.
Reads your GPU and runs the biggest Whisper model it can handle.
Map a keyword to longer text — say the shortcut, get the whole thing typed.
Optionally strips the “um”s and “uh”s as you talk, so you never backspace them.
Email, code, chat, docs — paste-first with a typing fallback.
Lives in the tray, starts at login, stays out of your way.
The popular apps are good — but most send your voice to the cloud, charge monthly, or lock the good stuff behind a paywall. Axon trades extras for free, offline, and no account.
| Axon Voice | Wispr Flow | Superwhisper | FluidVox | |
|---|---|---|---|---|
| Price | Free | $15/mo | Paid plan | $2.99/mo · $39 once |
| Free forever | Yes | Trial only | Trial only | Trial only |
| Account / sign-up | None | Required | Required | Required |
| API key needed | No | No | For AI extras | For AI extras |
| Works offline | Yes | No — cloud | Yes | Local tier |
| Voice stays on device | Always | No | Yes | Local tier |
Competitor pricing and offline status as of June 2026, from each vendor's own site. Wispr Flow runs in the cloud, so your audio is uploaded to be transcribed.
Axon does fewer things than the paid apps. No AI grammar rewriting, no “voice agent” that sends your emails, no 99-language cloud models.
That's the point. It's dictation that's free, private, and dependable — built by one person, for friends, family, and anyone who'd rather talk than type.
Free download. Install, hold a key, start talking. The first launch grabs the right Whisper model for your machine automatically.
↓ Download for WindowsAxon isn't signed with a corporate certificate yet — so Windows flags it as "Unknown publisher." It's not a virus. Here's how to get past it in two clicks:
When you see "Windows protected your PC", click More info (small text under the warning)
A "Run anyway" button appears at the bottom. Click it. That's it.
100% local · zero telemetry · open source on GitHub