VoiceGem vs Willow Voice: one Mac licence against a monthly seat (2026)
Willow Voice sells an unlimited free tier on a smaller model and charges $144 a year for the stronger one, across Mac, Windows and iOS, not the Mac alone. VoiceGem sells a macOS licence outright from $25, transcribes on the device, and converts spoken symbols into code with a fixed table rather than a model.
Willow Voice Pro is $144 a year for one seat, and enforced zero data retention only starts at the Business tier at $28 per user per month. (willowvoice.com/pricing, checked September 4, 2026)
Competitor pricing and features last verified September 4, 2026. Every claim links to its source.
| Willow Voice | VoiceGem | |
|---|---|---|
| Transcription on the devicesource | No | Yes |
| Code mode | Unverified | the Deterministic Pass |
| One-time licencesource | No | $25-$159, one-time |
| Free tiersource | Unlimited use of the smaller Frontier Mini model, limited personalization, 20 Scribe uses per week | 15 minutes of audio a day |
| Platformssource | macOS, Windows, iOS | macOS 14.4+ only, no Windows or iOS build |
| Documented data exportsource | Unverified | Import and export of settings and history |
Choose Willow Voice if…
- You want a free tier with no word ceiling at all — Willow Voice's free plan is unlimited on its Frontier Mini model, and VoiceGem's free tier is capped at 15 minutes of audio a day.
- You need Windows or iOS coverage, which VoiceGem does not provide in any form.
- You need enforced zero data retention with SSO for an employer, which Willow Voice sells at the Business tier and VoiceGem does not supply as a certification of any kind.
Editing models against the Deterministic Pass
Willow Voice's product is an editing layer: speech in, cleaned prose out, with model quality as the difference between its free and paid tiers. VoiceGem's Developer Mode replaces that layer inside an editor with the Deterministic Pass, a pure function with no model and no network call.
The Deterministic Pass matches spoken symbol phrases longest-first over three, two and one word, applies casing commands to the words that follow, and copies everything it does not recognise through unchanged. The result is that dictating the same signature twice produces the same line twice.
For prose, Willow Voice's approach is the better one and VoiceGem makes no argument otherwise. The claim is narrow and it is about code.
What switching between Willow Voice and VoiceGem involves
Willow Voice documents no export for a personal dictionary or history on its public pages, so the word list is retyped. Twenty minutes covers most people's list, and the subscription stops billing at the end of the term.
VoiceGem stores its dictionary on the Mac and exports its settings and history, so the rebuild is a one-time cost. Going the other way means giving up on-device processing, which Willow Voice does not offer as an option at any tier.
Willow Voice: best for, pricing reality, and what makes people leave
Best for people who dictate on more than one machine and want to evaluate without a word budget. Pricing reality: free with unlimited words on Frontier Mini and 20 Scribe uses a week, Pro at $15 a month or $12 on annual billing, Business at $35 a month or $28 annually.
The one thing Willow Voice does better than VoiceGem: an unlimited free tier and coverage of three operating systems from one account.
The one thing that makes people leave: transcription happens off the machine — Willow Voice's privacy policy says audio “may be processed transiently by Willow and its service providers” — and organisation-wide enforcement of privacy mode starts at the Business tier. Verdict: a good cross-platform subscription whose strongest guarantees are priced for organisations rather than individuals.
VoiceGem: best for, pricing reality, and what makes people leave
Best for a Mac developer who dictates into an editor and wants the audio to stay on the machine. Pricing reality: $25, $39, $49 and $159 for one, two, three and ten Macs, once, with a 14-day refund and 15 free minutes of recorded audio a day.
The one thing VoiceGem does better than Willow Voice: on-device transcription by default plus deterministic conversion of spoken symbols into source code, with presets for seven editors.
The one thing that makes people leave: the free tier is time-capped rather than unlimited, and VoiceGem is macOS 14.4 or later with no other platform. Verdict: cheaper after four months and unusable for anyone who needs a second operating system.
Which one to choose between VoiceGem and Willow Voice
Choose Willow Voice if you dictate across devices, if an unlimited free tier matters more than model choice, or if your employer requires contractual data-retention terms. VoiceGem cannot meet the last of those at all.
Choose VoiceGem if the work is code on a Mac and the audio must not leave it. Three years of Willow Voice Pro is $432 for one seat against $25 once, but that arithmetic only matters once the platform question is settled.
What people ask before switching
“My employer's code cannot go to a third-party server, so cloud dictation is out.”
VoiceGem's default engine runs on the Mac and the Deterministic Pass never leaves it. Cloud AI providers are opt-in and use your own key, held in the macOS Keychain. Nothing is sent anywhere unless you switch it on and supply the key yourself.
Which models run locally →“Fifteen minutes a day free is a trial, not a free tier.”
Fair, up to a point. It is 15 minutes of recorded audio per day, every day, with no expiry and no feature restrictions — Developer Mode and the local models are included. Call it a generous trial with an honest cap; it is not a two-week countdown.
The pricing page, including the cap →“When the connection is poor it stalls, spirals, or loses what I dictated.”
A hosted dictation service has a round trip to fail. VoiceGem's default engine runs on the Mac and the Deterministic Pass is a pure function, so there is no request to time out and no queue to spiral — dictation behaves the same on hotel wifi as at a desk.
Which models run on the device →Also comparing: what else does this job as well as Willow Voice, what Willow Voice actually costs, or moving off Willow Voice without losing anything.
Common questions
No. Willow Voice's privacy policy says audio “may be processed transiently by Willow and its service providers”. VoiceGem's default engine is on-device Parakeet, with eight Whisper variants and Apple Speech available, and nothing is transmitted by default.
Willow Voice's, on volume: unlimited words on its smaller model against VoiceGem's 15 minutes of recorded audio a day. VoiceGem's free tier withholds no features, so Developer Mode and the local models are included in it.
Willow Voice Pro at $12 per user per month on annual billing is $432 over three years. A VoiceGem licence is $25 once for one Mac, or $159 once for ten. Neither figure includes tax.
Willow Voice publishes no fixed spoken-symbol table and no deterministic code mode. VoiceGem converts 78 spoken phrases into 40 symbols and 7 casing phrases into identifiers with a pure function that has unit tests behind it.
VoiceGem retains nothing to begin with: the default engine runs on the Mac and session metrics stay in local storage. What it cannot offer is a contractual retention guarantee or a compliance attestation, which is what a Business plan buys.
Yes. Assign separate shortcuts, keep Willow Voice for the phone and Windows work VoiceGem will not do, and run Developer Mode in your Mac editor. The free tiers on both sides make a two-week trial cost nothing.
$25 once against $144 a year — and 15 free minutes a day to test it in your editor.
VoiceGem is released under GPL-3.0, a derivative of VoiceInk by Prakash Joshi Pax.