The Aqua Voice alternative for Windows where nothing leaves your PC

Published 5 September 2026 · updated 14 October 2026

On a Mac? Aqua Voice, or dictation that works with the Wi-Fi off

Aqua Voice runs on Windows, and its Windows page says the two things you should know before you install it: "Aqua processes audio in the cloud, where the Avalon model runs; there is no offline mode", and your first 1,000 words are free. After that it is a subscription, and every word you dictate is a recording sent to a server. Deixis is the other way to dictate on Windows. Hold a key, speak, and the words are typed where your cursor is, with the speech recognition happening inside the app on your own PC. No audio leaves. No account. Free, for everything that runs on your machine, with no word count. And it runs on your Mac and your phone too, with the same recognition on the device.

Aqua is fast and it is polished, and where it does something Deixis does not, more languages, a mode that shows your words as you speak, a voice command that sends the message for you, this page says so first. But the two tools disagree about one thing at the root, and it is the thing you cannot change with a setting: whether your voice has to go somewhere to become text. Aqua's pages say it does. Deixis was built so that it never does.

If you want dictation on Windows that is private because of how it is made rather than because of a promise, and free without a meter, read on.

Why switch

What you gain, and what you give up

Why Deixis

One tool, on every device you own. Windows, Mac and Android, and an iPhone keyboard in beta, with the same on-device speech recognition on each; your dictation audio never leaves any of them. On the iPhone it is a keyboard: switch to it, hold the microphone key, speak, and the text lands in the field. On Android it is a floating microphone bubble that dictates into any app, and the phone records and transcribes meetings on the handset as they run. Aqua offers Mac, Windows and iPhone, and its pricing page calls Android waitlist-only.
Your voice never leaves your PC. Aqua's Windows page says it processes audio in the cloud and has no offline mode; its FAQ says cloud dictation needs a connection. Deixis recognises speech inside the app on your own machine, under every setting, always. Pull the network cable and dictation keeps working, because it never used the network.
Nothing of yours is kept to train anything. Aqua's privacy policy says that with Privacy Mode off, it may store your transcripts on its servers to improve the product, and its own site summary says off is the default; you switch Privacy Mode on to stop that. In Deixis there is nothing to switch on: the transcript is written to a file on your PC and nowhere else, and the only copy is yours.
Free with no meter, and no account to start. Aqua's free Starter plan is 1,000 words, which its Windows page puts at about seven minutes of talking; its quick start is install and sign in, so the account comes first, and the meter runs from the first word. Everything local in Deixis is free forever and unmetered: dictation, hands-free notes, both speech models, the dictionary and all five vocabulary packs, screen capture and annotation, screen recording, and meeting recording with transcription on your PC. None of it asks you to sign in. The only time Deixis wants an email address is for the cloud extras, starting with the seven-day trial, and even then there is no card and no password.
It knows which field you are in without reading your screen. Aqua's home page says your screen is its dictionary, and its FAQ says that screen reading, Deep Context, stays off until you enable it. Deixis never reads your screen to dictate. It sends the name of the app and the field, so a commit message, a chat reply and a ticket are punctuated differently, and never the window title, never a screenshot, never the text on the page.
Talk and point at the same time. Free mode lets you narrate for as long as you like while grabbing a screen region, picking a button on screen, or taking the text you have selected. What you said and what you pointed at come out as one note, ready to paste into a bug report or a handover.
Screenshots, screen recording and meetings with the people in them named, all in the same key. Freeze the screen, drag a region, add arrows, boxes, numbered steps or a pixelated patch over anything private, and save. Record the screen, pause it from its bar, and stop it with the same key. Record a meeting, your side and theirs, and get a transcript written on your PC with the people in it named: Deixis tells the remote voices apart, you give each one a name, and from then on it recognises that person in later meetings. The voiceprint behind that is a mathematical fingerprint that stays on your machine, is never uploaded, and can be deleted per person. Aqua dictates; it does none of this.
Nothing about you goes out, and everything of yours stays as a file. Aqua's privacy policy describes session metadata, technical data, an advertising identifier, app-install attribution, and conversion measurement that can send a hashed version of your email address, with the sign-up or purchase it belongs to, to the advertising platforms it names. Deixis sends nothing about you at all: no analytics, no crash reports, no identifiers. What it keeps, notes, transcripts, screenshots, recordings and settings, sits in one folder on your PC as ordinary files. Back it up, search it, uninstall Deixis tomorrow and it is all still there and still readable.

Where Aqua Voice still wins

Forty-nine languages, with auto-detect. Aqua's Windows page lists 49 languages, and its language guide says it can switch between them automatically. Deixis ships English speech models only.
Words appear as you speak. Realtime Mode, on the Max plan, shows your words in Aqua's toolbar while you talk; Aqua's own site summary adds that this mode streams your audio to a third-party speech provider. Deixis types the finished sentence when you release the key.
A voice command that sends the message. Aqua's "Send it" voice command, also on Max, sends the message for you. Deixis puts the text at your cursor and leaves the sending to you.
Custom Instructions. On Pro and above you can tell Aqua how you want your text written and it follows the instruction. Deixis adapts to the app and the field it is typing into on its own, and takes transforms you say out loud, but it has no standing instruction you write once.
An API for its speech model. Developers can call Aqua's Avalon model directly. Deixis has no speech service to call; its extensibility is buttons that run your scripts, webhooks that post finished notes to URLs you own, and a local connection for programs on the same PC.
Cheaper for students, and team controls. A 70 per cent student discount, and Team and Business plans with centralised billing, an organisation-wide Privacy Mode and single sign-on. Deixis has teams and seats, but no student price and no sign-on integration.

What you get

More than dictation. One key, and everything behind it.

Hold the key and press a letter: free mode, a grab, a meeting, a recording, a translation, your own buttons. All of it on your own device, all of it free, none of it behind an account.

The Deixis panel tethered to the caret in a code editor, listening — Dictation, where your cursor is

Hold a key, speak, release. Clean, punctuated text lands in whatever has focus, on your PC, your Mac and your phone, with your clipboard handed back.

A finished free-mode note with two captured references interleaved with the narration — Free mode: talk and point

Narrate for as long as you like while you grab a region, pick a button on screen or take a selection. One note, what you said interleaved with what you pointed at, pasted where you finish and saved as Markdown.

A frozen screen with a dashed selection, numbered steps, an arrow and a pixelated patch — Screen grabs, annotated

Freeze every screen at the keypress, drag a region, add arrows, boxes, text and numbered steps, and pixelate what is private. Saved as a PNG that stays editable.

The page Deixis writes for a shared recording: a short link, a title, a summary and the video — Screen recording, shared as a link

Record the screen with your narration. The recording writes itself up with a title, chapters and a summary, and from your Deixis account it shares as a link anyone can open without one. The Loom you can stop paying for.

A meeting transcript with timestamps and named speakers, and a level meter — Meetings, no bot, names included

Your microphone and the other side, transcribed on your machine while the call runs. Deixis tells the voices apart and remembers the ones you name, and never uploads a voice.

A custom button definition, a webhook delivery and the vocabulary packs — Your own buttons and workflows

Add a button to the palette: it gets the note on stdin and types back whatever it prints. Send finished notes to any URL you own. Chain steps: read the text off a screenshot, translate it, summarise it, post it.

Side by side

Aqua Voice and Deixis, side by side

Row	Deixis	Aqua Voice
Runs on	Windows 10 and 11 (64-bit, on an Intel or AMD processor with AVX2: 2013 and later; no build for Windows on ARM), Mac (Apple Silicon) and Android (a dictation bubble in any app, plus meetings), with an iPhone keyboard in beta on TestFlight and not yet in the App Store. Meetings run on Windows and Android. One engine on all of them, and dictation audio never leaves the device.	macOS (one universal download for Apple Silicon and Intel), Windows 10 and 11, iOS
Speech recognition happens	On your own device, inside the app, on every platform. Dictation audio never leaves it; a meeting's, only if you turn on cloud processing on Android.	In the cloud
AI cleanup happens	On your device, with no provider at all. Or, if you want more, Deixis's cloud, your own Ollama, or your own key. Your choice, in one setting.	In the cloud
Leaves your machine	Text, and only once you switch a cloud feature on, plus, on Windows, one key you press to check your own Claude plan, and, on Android, a meeting's audio if you turn on cloud processing for meetings. Never a dictation's audio, never screenshots, never window titles.	Audio, text
Account needed	None, ever, for anything that runs on your device. The cloud extras have a seven-day trial you start with an email code. No card.	Required
Price	Free forever for everything local, meetings included. Pro is $49 a year for the cloud extras with nothing to set up.	Free tier: 1,000 words on the Starter plan, no card; its Mac and Windows pages call that about seven minutes of talking; $10 a month; $96 a year
Your transcripts live	In plain files, in one folder you own and can back up. Uninstall Deixis and every note is still readable.	On your device and in the cloud
Make it yours	Add your own buttons that run a script, send finished notes to any URL you own, and switch on vocabulary packs for your field.	A custom dictionary and replacements, Custom Instructions on Pro, a Send it voice command and Realtime Mode on Max, and an API for its Avalon speech model

Aqua Voice facts from home, aquavoice.com, its privacy policy, windows, faq, terms, download, history, languages, mac, llms and realtime, read on 27 September 2026. “Not documented” means those pages did not say.

Aqua's Mac and Windows pages say it processes audio in the cloud, where its Avalon model runs, and that there is no offline mode. With Privacy Mode off, which its own site summary says is the default, transcripts may be stored on its servers to improve the product. Realtime mode, on its Max plan, streams your audio to a third-party speech provider.

What leaves your machine, and what leaves theirs

Deixis publishes every byte it sends, as a table, because there is so little of it. Here is that table beside what Aqua Voice's own pages say.

What	Deixis	Aqua Voice
Audio	A dictation's, never: recognition happens inside the app on your device, and the recording, the model and the transcript stay there. A meeting's, only if you turn on Deixis cloud processing for meetings on Android: its audio, in pieces, each its own request, and Deixis stores and logs none of itapi.deixis-voice.com, then Whisper on Cloudflare Workers AI · only on Android, after a meeting ends, with cloud processing switched on in Settings › Meetings — on Wi-Fi, or on mobile data only if you choose	Your voice, to be turned into text: its Mac page says Aqua processes audio in the cloud, where its Avalon model runs, and that there is no offline mode. Aqua's site summary adds that Realtime mode, on the Max plan, streams audio to a third-party realtime speech provider. Its history guide says recent sessions' audio is also kept on your device and deleted after three days.Aqua's cloud, and in Realtime mode a third-party realtime speech provider · every dictation
The dictated text	The sentence you just dictated, your vocabulary, and your previous sentence in that fieldthe endpoint you chose · only with a cloud endpoint set, and by default only when the local rules are not enough	The transcript is produced in the cloud. The privacy policy says that with Privacy Mode off, which Aqua's site summary gives as the default, Aqua may store transcript data on its servers to improve the product; with it on, transcripts are not collected, though session metadata may still be, and the home page says nothing is then stored on its servers.Aqua's servers · every dictation
Context around the text	The names of the app and the field the text is going into. Never the window title, never a screenshot.the same endpoint · with that request	Deep Context, its screen reading, is something you enable, per the FAQ; the home page says Aqua understands what is on your screen, from code syntax to everyday text, and Aqua's site summary calls it a client-side engine whose data is never stored. The pages do not say what screen content, if any, is sent with a dictation.not stated · only with Deep Context on
Telemetry and analytics	None. No account, no telemetry, no analytics, no crash reports.	Session metadata (timestamps, device type, performance metrics) and technical data (IP address, browser type, operating system). For advertising measurement and app-install attribution, a first-party identifier, and measurement data that may include a hashed version of your email address, sent with the event to the advertising platforms the policy names.Aqua, its attribution provider and the named advertising platforms · not stated
Update check	Nothing but the request; the version check happens on your PCapi.deixis-voice.com · every launch, and when you ask in Settings	not documented
Model download	Nothing but the requesthuggingface.co · once, when you press Download on first run	not documented
Account and sign-in	An email address, only if you choose to start the trial, plus the partner tag from a partner download link and the first partner website that connected to Deixis on this computer, if there are anythe Deixis endpoint · when you start the trial	Name, email address, billing address and payment informationAqua · when you create the account and when you pay
Integrations	The finished note, and for a screenshot its file name, never the imagea URL you added yourself · on finish, only for destinations you added	not documented

The same table for every tool we have measured

Before you switch

Who should not switch

If you dictate in a language other than English, stay with Aqua; Deixis's speech models are English models. If your PC is older than 2013, or runs Windows on ARM, Deixis is not for it: it needs an Intel or AMD processor with AVX2. And if your desk is a Linux box, Deixis is not there. And if watching the words appear while you are still talking is what you love about Aqua, Deixis types when you let go of the key, not before; that is a real difference in feel and no setting changes it.

Making the switch

Five minutes from installed to dictating. Here is what maps to what.

The dictation key

Aqua's Windows page describes it as hold the key, talk, release. Deixis is the same gesture with a different key: hold Ctrl+Win, speak, release, and the text lands at your cursor with the clipboard put back the way you left it. Double-tap the same key instead and it keeps listening until you double-tap again, for the paragraph you have not finished thinking yet. You can change the key in Settings if another one suits your hands better.

The first run

Aqua signs you in and counts your first 1,000 words. Deixis shows one Download button. Press it, wait for a file about the size of a short video, and you are dictating. That is the speech model, fetched once and kept on your PC. Settings offers a second, larger model that is more accurate and slower to decode, and it swaps in without a restart.

Your dictionary and replacements

Aqua has a custom dictionary and replacements. Deixis has both too, and you can add a word without leaving what you are doing: select it anywhere, press Ctrl+Win+D, done. Five vocabulary packs ship in the box, for software, finance, business, medicine and law, and switching one on takes a click. A word it keeps mishearing can be corrected once and fixed forever, and the corrections are unlimited.

Custom Instructions, and what replaces them

Aqua lets you write instructions for how your text should come out. Deixis works from where the text is going instead: it sends the name of the app and the field along with your words, so the cleanup already knows a terminal from an email. You can also ask for a transform out loud, "make this a bullet list", and it will. What it does not have is a standing instruction you write once.

Privacy Mode, and what replaces it

Aqua's Privacy Mode decides whether its servers keep your transcripts. In Deixis there is no such setting because nothing is sent to be kept. The one choice that matters is whether AI cleanup uses the network, and it is in Settings → Cleanup. Out of the box, Deixis tidies your dictation on your PC with no provider and no key. If you want a language model to polish your sentences, choose between Deixis's cloud, a model running on your own machine, or your own key. Whichever you pick, the page linked at the bottom of this one says exactly what each choice sends, and you also choose how often: only when the local rules are not enough, every time, or never.

History

Aqua's History tab is an on-device record of your sessions, with the audio deleted after three days; by its privacy policy, transcripts may also be stored on its servers unless Privacy Mode is on. Deixis keeps your dictation history in the Library and every note, transcript and recording as a file in one folder on your PC, for as long as you like.

Meetings, screenshots, recordings

Ctrl+Win+M records a meeting, both your voice and theirs, and transcribes it on your PC. The remote voices come out told apart, and you name them: give each voice a name once, and the next time that person is in a meeting Deixis recognises them and labels them. The voiceprint it uses is a mathematical fingerprint that stays on your machine, is never uploaded, and can be deleted per person. Ctrl+Win+S grabs and annotates a screenshot. Ctrl+Win+V records the screen to a video file. All of it lands in one folder as files you can open with anything.

Before you ask

Questions people search for

Does Aqua Voice work on Windows?

Yes. Aqua's Windows page says it supports Windows 10 and 11, and that you dictate into anything you can place a cursor in. It also says that Aqua processes audio in the cloud and has no offline mode, and its FAQ adds that cloud dictation needs a connection. Deixis is the Windows tool that does the recognition on your own PC instead, offline; it runs on the Mac and on your phone the same way.

Does Aqua Voice work offline?

Its Windows page says no: Aqua processes audio in the cloud, so it needs an internet connection, and there is no offline mode. Its FAQ is a little softer, saying offline availability varies by platform, app version and mode, but that cloud dictation needs a connection. Deixis works offline in every mode, because the speech recognition never used the network in the first place; the only calls a fresh install makes on its own are an update check and, when you open the Settings window, a check for an announcement from us, and neither carries anything but the request.

Is Aqua Voice free?

The Starter plan gives you 1,000 words free with no card; after that, Pro was $10 a month, or $8 a month billed annually, and Max $30 a month or $24 billed annually, when we checked. Deixis is free forever for everything on your PC, with no word count and no account. Pro, at $49 a year or $10 a month, adds the managed cloud extras with nothing to set up.

Does Aqua Voice store my transcripts?

By default it may. Its privacy policy says that with Privacy Mode off, transcript data may be stored on its servers to improve the product, and its own site summary says off is the default; its home page says that with Privacy Mode on nothing is stored on its servers. The audio still travels to its cloud to be transcribed either way. Deixis has no servers in the path: recognition happens inside the app on your PC and the transcript is a file in your folder.

Does Deixis type while I am still talking?

No. Deixis types the finished sentence when you release the key, cleaned and punctuated. If you want to keep talking, double-tap the key for hands-free mode and it keeps listening until you double-tap again, then pastes the whole note at your cursor.

Does Deixis upload my audio?

Not a dictation's, under any setting. Speech recognition happens inside the app on your PC, and on your Mac and your phone it happens on the device the same way. A meeting's audio stays on the device too, unless you turn on Deixis cloud processing for meetings on Android, which sends a finished meeting's audio to be transcribed in the cloud, and Deixis keeps none of it. Otherwise the only content that can ever leave is text, and only once you switch a cloud feature on. One key press is the exception that carries no content at all: Claude usage reads the sign-in Claude Code has already stored on your machine and asks Anthropic what is left of that plan. The full list, every byte, is one page long and linked at the bottom of this one.

Keep reading

Hold a key. Speak. Keep working.

Deixis is free, needs no account, and runs on your own device. Everything on this page is in the current release.

Download Deixis for Windows or Mac Every byte that leaves

v1.3.2 · 64‑bit Windows 10/11 on Intel or AMD · macOS 13 on Apple silicon · 51 MB · nothing else to install · Android on Google Play

Facts about Aqua Voice checked on 27 September 2026 against aquavoice.com. Deixis facts are from its documentation for the current release. Something out of date? Tell us.

The Aqua Voice alternative for Windows where nothing leaves your PC