Hello.

Speak. Write. Do.

Ready

Ready when you are

Press in any app — or tap the button.

✎ Voice edit
· TestChange

Your stats

Your Voices

Speak naturally. LexVoca writes it in the style that fits.

AI writing configured

How should LexVoca write?

You say · example
LexVoca writes

Your styles

Which style in which app

Every other app in Match the app: no style rewrite.

◇ Checks for changed details. LexVoca checks for some changes to numbers, names and negation. A failed check keeps the earlier text. Review important wording.

◷ Review your words. History shows recognised text alongside the result where available, until entries are deleted.

☁ Written by . Styles apply to eligible dictation; protected inputs are skipped.

Copy earlier Voice edit styles into Your Voices

Copying creates a draft. Existing Voice edit styles stay separate.

History

Saved dictations and activity, newest first. History stays on this PC.

Say "voca again" to type the newest one again.

Hover or focus a line for Copy, Type it again, Read aloud and Delete. Deleting one entry offers Undo for five seconds. Speed details are in Details.

Dictionary

Words LexVoca should spell your way. Used in every dictation.

Learning

From AI tidyingWhen the same word keeps being fixed.
When I fix a word myselfReads that one text box briefly. Privacy

Words for your line of work

Preview words

Your words

→

Move your words, trigger phrases and app styles to another computer.

Tasks & notes

Say it and forget it. LexVoca works out whether it's a task or a note, and when it's due.
Use the microphone or type above.

Tasks

Notes

When a task is due, a reminder pops up over whatever you're doing — even when LexVoca is minimised or in the tray. Repeating tasks remind you each time.

●
Your reminder
Due now · preview only
In 10 minDone

Trigger phrases

Say a short phrase and LexVoca types a whole saved block — a signature, a link, a standard reply.

New trigger phrase

Two or three words you wouldn't say by accident.

Your trigger phrases

Start from an idea

"email signature"Your sign-off, name and title.
"standup update"Yesterday, today, blockers — ready to fill in.
"out of office"A polite away message with your dates.

Commands

In a terminal, say what you want. LexVoca types the command for you to review.

Try a command — nothing is typed anywhere

Spell it

For short commands.

“L S dash L A” → ls -la

Say the words

For longer ones.

“get child item” → Get-ChildItem

Just ask

In plain words.

“show disk space” → df -hT

Built-in commands are checked against the allowed list, including navigation and supported folder moves. Saved phrases can contain commands that change things; review them before running. Auto-enter settings

Which windows get commands

PowerShell and Command Prompt are recognised by their title. For PuTTY or SSH, LexVoca asks and remembers the program or user@host. Say “normal text” in a terminal to switch that window back to words.

Your command phrases

Words after the phrase are added: “list details slash tmp” types ls -la /tmp.

Words LexVoca hears wrongly

What LexVoca understands

Feedback

Tell us what went wrong, what you would like next, or a command LexVoca got wrong. It goes straight to the people building LexVoca - no website, no signing in.
Exactly what will be sent

            

Your feedback

Kept on this computer, with whether each one has reached us. Anything not sent yet goes automatically the next time LexVoca can reach the service.

Launch

Open a program by saying its name: launch Word, open Calculator, switch to Outlook. In a terminal, use launch explicitly so shell requests take priority.

How it works

LexVoca opens programs discovered in your Start menu or local launchers you explicitly add under Your own apps. Only add programs and shortcuts you trust. Recognition mistakes can open the wrong app.

For discovered programs, LexVoca tries to bring an existing window forward. Custom launchers start the selected file. Switch to only brings an existing window forward.

When several programs match, use your dictation shortcut again and say a listed number or name, or click a choice. Say cancel to stop. Outside this choice step, unrecognised program requests remain ordinary dictation in text mode.

Turn this off to have these words typed like any other dictation.
Shows the program it heard and waits for you to confirm. Slower, but nothing opens by surprise.

Words that start a program

One per line. The first word you say has to be one of these, which is what keeps ordinary dictation from opening things.

Your own apps

Give a local program or shortcut a spoken name. Choose an .exe or .lnk file, without command-line arguments. These paths stay on this computer and are not synced.

Folders and desktops

Say “open Documents”, “open Downloads”, “open Pictures” or “show desktop”. Virtual-desktop switching is not available in this version; use Windows Task view to choose a desktop.

Use an existing local folder. These names open folders; they do not move or change files.

Programs LexVoca can open

Programs from your Start menu, with suggested categories you can change. Friendly names are alternative names you can say after "launch". Separate multiple names with commas.

Apps & websites

Most apps need nothing here. Use this when your words don't arrive properly in a window, such as a console in a web page or a remote session that ignores Ctrl+V.

Check a window

Press Check, then click into the app or page within 3 seconds. Nothing is typed or pasted.

Apps you've taught it

Also set from the tray menu, for the window in front.

Every other app: LexVoca decides automatically.

Websites

In Automatic mode a website rule wins over the browser's app setting.

How LexVoca decides

  1. A website rule

    The closest matching address wins.

  2. An app you've taught it

    Used when no website rule fits.

  3. Automatic

    LexVoca works it out from the window.

An explicitly forced method in Settings overrides these rules. Dictation and clipboard Superpaste have separate settings.

LexVoca reads only the address bar of supported browsers on this PC, never page contents or browsing history.

More detail

Edge, Chrome, Firefox, Brave, Opera and Vivaldi are supported where the address bar is accessible. Query strings and fragments are ignored. Hidden protocols are matched by host, port and path; conflicting matches or an unavailable address fall back to the app/default method. A //host/path rule accepts either HTTP or HTTPS.

Settings

Dictation

How LexVoca treats what you say

LexVoca can type your words, or turn them into command lines in terminals and consoles - typed at the prompt, never run by themselves. See the Commands page.

Commands

Windows only lets a program type into windows at its own level: an "Administrator:" terminal ignores keystrokes and mouse buttons from a normal LexVoca. This runs LexVoca as administrator through a scheduled task (one permission prompt to set up, none afterwards) and restarts it.
Only when the built-in grammar cannot make a command line: your connection's chat model suggests one, and LexVoca checks it against the same allowed-command list before typing. History labels those lines "AI suggested". Off means unrecognised words are simply not typed.
Presses Enter for supported basic checks and common folder navigation in other shells. Bash always requires draft review and manual Enter. Commands that may change things, AI suggestions and uncertain results always wait for you. If a line is left unsubmitted, Auto-enter pauses until you confirm the command line is empty. Off by default.
In a Bash terminal, supported requests such as "list hidden files here" become an exactly quoted draft. Review the draft and choose Type before it enters the terminal; LexVoca never presses Enter for Bash. If a place is unclear, choose an offered destination or give an explicit path. Unavailable context stays unavailable. Saved phrases and other Bash commands also require review before typing.

More options appear in the Technical profile.

Use the English spelling that suits you. Other languages stay as requested.

Language

Speak in one language and have LexVoca type another. Names, numbers and snippets are kept. Part of the NeuroDynamic AI Pro plan.
You say"Meet at 3 pm on Thursday" LexVoca typesMeet at 3 pm on Thursday.
Your own speech server translates while it listens, if it supports that. AI translation uses your NeuroDynamic account with any speech choice.

Shortcuts

Keyboard shortcuts

They work in any app. Press Change, then press the keys you want.

Records speech and types it where the cursor is.
Rewrites selected text using your spoken instruction. Blank turns the shortcut off.
Records a task, note or reminder in LexVoca instead of typing into another app. Blank turns the shortcut off.
Checks Windows reservations, LexVoca duplicates and family clashes. Defaults reset all four keyboard shortcuts.

Press or hold?

How the keyboard shortcuts start and stop dictation.

Mouse button

Easier than keys: most mice have two side buttons. Pick one to start and stop dictating.
Stops the app underneath also acting on the button - Notepad++ and many browsers map the side buttons to "previous/next tab", which jumps the page while you dictate. Ordinary clicks are untouched. Where you were typing also reserves the dictation button to preserve the caret. This option additionally reserves configured mouse buttons in pointer mode.
Where you were typing keeps the existing keyboard caret, wherever the pointer moves, and reserves your dictation button for LexVoca. Changing field or app during dictation may prevent insertion; your words remain in History. Window under the pointer uses the earlier target behaviour.
The mouse button's own choice, separate from the keyboard.

Typing & pasting

How your words get into the app

Some windows, like remote desktops and virtual machines, can't take a paste. LexVoca types into those instead. New dictations only.

Characters per second for typed input. 0 uses 40; lower values suit slower consoles. Host and guest keyboard layouts must match.

Superpaste: paste your clipboard anywhere

Types what you copied into windows that ignore Ctrl+V - terminals, remote sessions and virtual machines. Per-app and per-website choices live on the Apps & websites page.

Sends clipboard text to the focused app or console. Ordinary Ctrl+V is unchanged.
Sends copied text to the focused app. Must differ from the dictation mouse button.
Superpaste already works in every virtual machine from the mouse button, the tray menu or the delayed paste. A VirtualBox console captures the keyboard so completely that the Superpaste shortcut never reaches LexVoca; with this on, LexVoca watches for that one shortcut while a VirtualBox console is in front. Nothing else is read or kept.
Auto chooses keys for VMs and browsers. Delayed paste gives you three seconds to focus the destination. Escape stops typing.Not tested

Finishing touches

Brings the window you were in when recording started back to the front before inserting, so the text still lands there if you clicked away while speaking. Off inserts wherever you are when it finishes.
Restores your previous clipboard text after LexVoca pastes, unless you copied something new.
One space after dictated text, ready for your next sentence.
For dictating one sentence in several goes. Without this, every dictation ends in a full stop. With it, say "full stop" to finish the sentence; otherwise the next dictation carries straight on, with the capital letter dropped.

Hands-free voice

Say a phrase instead of pressing a key. LexVoca listens on this PC only for these phrases.

Listens locally while a console or browser is focused. Confidence checks reduce accidental triggers; speech stays on this PC. Uses the Windows default microphone; pauses during dictation.Off
Types whatever you last copied. At least two words.
Uses the destination's normal paste or typing method.
Starts recording with no key pressed, and stops by itself when you stop talking. With Auto-enter on, supported basic terminal commands can run by voice alone; other commands still need Enter.
During a dictation, saying only the paste or repeat phrase performs that action when supported. This does not enable background listening in other windows.
For hands-free dictation. Raise it if it keeps listening in a noisy room; lower it if it cuts you off.

Speech

Where your speech is turned into text

Each choice keeps its own speech settings. AI features use your NeuroDynamic account separately.

This choice is for speech only. AI features — tidying, Your Voices and voice edit — use your NeuroDynamic account whichever you pick.

Hosted speech access

Hosted authentication is separate from recording recovery. AI sign-in is in Settings › AI.

Unfinished recordings and requests

For every speech choice. Review saved recordings and AI requests here; History keeps recognised text separately.

History keeps recognised text. Saved recordings remain on this PC until recovered, deleted or expired.

Speech model and server

Saved keys are kept in Windows Credential Manager
Speech model offered by this server. Refresh loads its model list.
e.g. en - blank auto-detects
Speech: not tested Cleanup: not tested

AI

Your other NeuroDynamic apps, ClaraVoca and ITSpecta, use this sign-in too. Signing out here signs you out of all of them.

NeuroDynamic AI

Tidying, writing styles, voice edit and more. NeuroDynamic picks the right model for each job.

Punctuation, capitals and slips. It never rewrites, answers or acts on your words.
Suggests a command when LexVoca doesn't recognise one. You still press Enter.
Your VoicesWrites in the style that fits each app.
Voice editSelect text, press and say what to change.
Write my dictation in another language ProSpeak one language and LexVoca types another.

Try it

See the tidied result before you dictate. Nothing is typed.

Try it works once AI is available again.

LexVoca types

Instructions

Rewriting style per app

For named dictation styles, open . These existing instructions remain voice-edit only.

Extra instructions used only when you voice-edit selected text. Ordinary dictation keeps your wording and ignores these.

→

Appearance

Theme

The pill

The small bar that floats over your apps while you dictate.

Shows the pill, small, at the bottom of the screen when you are not dictating - with the buttons below on it. Off: it only appears while you dictate and for the message afterwards.
Small coloured buttons for the things you do most. Clicking one never takes the focus from the window you are working in.
Replaces action colours only. Symbols and labels identify each action too.
Keeps LexVoca's icon on the taskbar instead of inside the hidden "^" flyout. Windows only records an icon once it has appeared, so open the flyout once if this does not take effect.
Right-clicking the tray icon opens the plain Windows menu instead of LexVoca's styled one. Screen readers announce the standard menu fully.

Microphone and sounds

System default follows Windows when the next recording starts. A selected device is never silently replaced. The background listener uses the Windows default.
One quiet sound when LexVoca starts and stops listening, and when something needs you.
Choose a short tick, click or beep. Original keeps the earlier two-tone chirp.
LexVoca's own level; Windows volume still applies.

Backup & more

Troubleshooting

Diagnostic logs contain timings and safe outcome codes. Dictated text, recordings, History, settings and credentials are excluded from exports.

Back up your set-up

Keep a copy, or move LexVoca to another computer. Keys and passwords are never included.

Never included: keys and passwords, your server connections, where updates come from, and "Auto-enter" - set those up on each computer. Restoring adds to what you have; nothing is deleted.
Keep the file somewhere safe, or open it on another computer.
Kept with your account, so any computer you sign in on can restore it. Needs the hosted connection set up; it will say so if the service is not ready yet.

ClaraVoca, your reading companion

Not checked yet

Updates

Only signed updates from LexVoca's update share are installed.

How LexVoca works

Short answers, one topic at a time. Each section starts with the gist; open More detail only when you want it.

Your first dictation

Click where you want the words, press , speak, then press it again. Your words appear where your cursor is.

  1. Click into the place you want to write: an email, a document, a chat box.
  2. Press (the left Alt key). The pill appears at the bottom of the screen and listens.
  3. Speak normally, in whole sentences. Pausing to think is fine.
  4. Press again. With Type where I started on, LexVoca brings the original window forward before inserting. Keep the intended text field selected; changing fields can prevent insertion.
  • Prefer to hold the keys while you talk, like a walkie-talkie? Choose Hold while you speak in Settings, Shortcuts.
  • A spare mouse button can start and stop dictation too.
  • Recognised text is saved in History until deleted. If recognition did not finish, check Settings, Speech for a saved recording; recovery depends on what was saved and your retention setting.

Where to change it

About the name

LexVoca is said LEX-voh-kuh. Lex, from lexicon: words. Voca, from the Latin vocare: to call. Words, called: you speak, and your words appear in a supported text field.

  • Speak. Write. Do. Speak is the voice, write is the words, and do is the calling: the name acts.
  • ClaraVoca, the reading companion, shares the root. Clara, from clarus: clear. Words, made clear.
  • Always written with a capital V: LexVoca, ClaraVoca.

Where to change it

The main shortcuts

All of them use the left Alt key, so the right Alt key stays free for typing accents.

  • Inside LexVoca: CtrlK finds a page or action, Ctrl, opens Settings, F1 opens Help, Esc closes what's open, and Alt1 to Alt9 open the sidebar pages.
  • To fix the last dictation, use the tray menu.
  • Dictate
  • Voice-edit the selected text
  • Speak a task or note
  • Superpaste
  • NeuroDynamic menu: LexVoca, ClaraVoca and ITSpecta AltQ

Where to change it

Profiles, and Simple or Advanced

Your profile decides which features LexVoca uses. Simple or Advanced decides how much each screen shows.

  • Dictate: your voice to text, everywhere: dictation, AI tidying, Your Voices, voice edit, tasks, trigger phrases.
  • Assisted: everything in Dictate, plus opening apps by voice, hands-free phrases and reading aloud with ClaraVoca.
  • Technical: everything in Assisted, plus commands in terminals, virtual machines and consoles.
  • Simple shows the main controls. Advanced shows additional settings for your current profile. Switch at the bottom of the sidebar.
  • What a profile hides is also switched off, so nothing you can't see acts on your words. LexVoca remembers how you had it and gives it back if you switch again.
  • Cards with tracked Advanced settings show a custom setting count when those values differ from their defaults. Use Show to inspect them.
  • Your dictionary, history and connection are the same in every profile.

Where to change it

The pill: what it's telling you

The small bar at the bottom of the screen shows what LexVoca is doing while you work in another app.

  • Ear and moving bar: listening. The bar follows how loudly you speak.
  • Soft wave: turning speech into text, then tidying it.
  • Tick: the operation has finished or been saved. Read the message to see which; timing may also be shown.
  • “i”: a note, such as “No speech detected”. “!”: something went wrong, and where to recover your words.
  • Keep the pill on screen to have it rest at the bottom with quick-action buttons, such as Dictate and Paste in 3 seconds.

Where to change it

LexVoca can't hear me

Usually the wrong microphone is chosen, or Windows is blocking it. These checks fix most cases.

  1. Open Settings, Appearance, Microphone. System default uses the microphone Windows has set as default; if your headset isn't the default, pick it by name.
  2. Plugged it in after LexVoca started? Press Refresh next to the list.
  3. Check Windows lets apps use the microphone: Windows Settings, Privacy & security, Microphone, and allow desktop apps.
  4. Check the headset isn't muted: many have a mute switch or button on the cable.
  5. Try it: Settings, Speech, Start speech test.
  • “No speech detected” can mean the microphone is muted, the wrong device is selected, or the recording was too quiet or short. Check the selected device and try a full sentence.

Where to change it

Getting the best accuracy

A good microphone close to you, and speaking in whole sentences, make the biggest difference.

  • Best microphone: a headset, or a desk microphone a hand's width away. Laptop and webcam microphones pick up the room as much as you.
  • Speak at a steady, normal pace, in full sentences. LexVoca uses the words around a word to choose the right one.
  • Start speaking once the pill is listening; a soft sound plays if sounds are on.
  • Keep background noise down: music, fans and other voices all compete with you.
  • Turn on Tidy up what you say (Settings, AI) for punctuation, capitals and small slips.
  • Long dictation? Pause between ideas rather than mid-sentence.
  • Home's Noise control changes the level used to detect pauses in hands-free dictation. It does not remove background noise. Raise it if recording keeps listening; lower it if it cuts you off.

Where to change it

Names, accents and unusual words

Accents, names, abbreviations and slang can be misheard. Dictionary entries and vocabulary lists can help, but check important wording.

  • Learn from AI tidying: when enabled, repeated corrections can become Dictionary entries after LexVoca's checks.
  • Learn from your corrections: when enabled in Dictionary, supported edits made soon after dictation can suggest or teach a correction. Not every edit can be detected.
  • Add it yourself: in the Dictionary, type what LexVoca hears and how to write it, for example “it spectre” becomes “ITSpecta”.
  • Abbreviations: add how they sound, such as “A P I” becomes “API”.
  • Your line of work: add a ready-made word list, such as IT support, health, legal or finance.

Where to change it

My words didn't appear

If a transcript was saved, recover it from History. If recognition did not finish, check for an unfinished recording in Settings, Speech.

  1. Open History and press Copy or Type it again on the entry.
  2. If it happens in one app or web page, Technical users can use Check a window on Apps & websites and choose another method.
  3. Windows opened “as administrator” ignore normal apps. Technical users can turn on Work in administrator windows in Settings, Dictation.
  • Make sure a text box has the cursor before you start. With Type where I started on, LexVoca goes back to it for you.

Where to change it

Nothing happens when I press the shortcut

Another program may be using the same keys, or the right Alt key was used.

  1. Use the left Alt key. On many keyboards the right one is AltGr and types characters.
  2. Settings, Shortcuts: press Scan conflicts. If a shortcut is in use, choose another.
  3. Some graphics-card overlays use Alt + Z. Change the overlay's shortcut, or LexVoca's.
  4. Inside a virtual machine the guest may grab the keyboard. Use a mouse button instead.

Where to change it

The spelling is American, or British

Choose the English spelling you want in Settings, Dictation, Spelling.

  • Follow Windows uses your Windows region. British English or American English fixes it either way.
  • If an old Dictionary entry flips a word back, remove it in the Dictionary.

Where to change it

“Speech could not finish”

The speech request did not finish. Check the message for a connection, sign-in, service or on-PC model problem.

  • Open Settings, Speech and check the selected connection. Test connection checks access; use Start speech test to test an actual recording.
  • Look under Unfinished recordings for saved items. Available actions depend on the item; recovered text goes to History and is not automatically typed. Recordings may have expired or been deleted.

Where to change it

A reminder didn't pop up

LexVoca must be running and the PC awake to show a reminder at its due time. Minimised or in the tray is fine; quitting LexVoca stops reminders.

  • Open the task with Edit and check its day and time. Check Windows notifications too if no notification appears.

Where to change it

AI says paused, or asks me to sign in

AI and speech have separate availability. AI tidying can fall back to a recognised transcript, but hosted speech still needs a valid sign-in and working service.

  • Sign in to use AI features: sign in with your NeuroDynamic account in Settings, AI.
  • AI running low or AI paused until a date: your plan's AI use resets on that date.
  • Confirm your email: open the link NeuroDynamic sent you.
  • Service paused: check the service status. If speech also failed, look for a saved recording in Settings, Speech.
  • Manage account opens the website; being signed in there does not prove the app is connected. Use Check AI access or reconnect in Settings, AI when offered.
  • AI switched off: turn it on in Settings, AI if you want features that require it. Speech on this PC does not provide a local AI writing service.

Where to change it

LexVoca typed a command when I wanted words

In a terminal, Automatic mode turns speech into commands. You can tell it otherwise.

  • Say “normal text” in that terminal, or set it under Which windows get commands on the Commands page.
  • Or choose Always text in Settings, Dictation.

Where to change it

AI with NeuroDynamic

Tidying, Your Voices, voice edit and translation use your NeuroDynamic account, whichever way your speech is turned into text.

  • Tidy up what you say: aims to fix punctuation, capitals and small slips while keeping your meaning. Review important wording.
  • Your Voices: writes in the style that suits each app.
  • Voice edit: select text, press and say what to change.
  • Write my dictation in another language is part of the Pro plan.
  • NeuroDynamic picks the right AI model for each job, so there is nothing to set up or tune.
  • When AI is on, the recognised text needed for that feature is sent to NeuroDynamic. With speech on this PC, your audio stays on this PC.
  • If AI tidying is unavailable, LexVoca can keep the recognised transcript instead. Voice edit and other AI-only actions need AI access; their messages explain when they are unavailable.

Where to change it

Your Voices

Speak naturally and LexVoca writes it in a style: a work email, a quick message, a report, or your own.

  • Just as I say: no style rewrite. Tidying still depends on Tidy up what you say being enabled and available.
  • Match the app: each app gets its style, such as Work email in Outlook.
  • One style everywhere: one style for eligible dictation. Protected inputs, including terminals and remote consoles, are skipped.
  • Make your own style by describing it or from samples of your writing, and try it before you use it.
  • Learning examples are sent only after you agree and are not saved by LexVoca. A successful preview is saved with the style when you press Save, and may be included in a writing-style backup. Your text service may retain sent text under its own rules.
  • Terminals and remote consoles are never rewritten into prose.
  • LexVoca checks for some changes to numbers, names and negation; a failed check keeps the earlier text. These checks cannot guarantee that every change in meaning is detected.

Where to change it

Voice edit

Select some text, press and say what to change: “make this friendlier”, “shorten it”, “turn it into bullet points”.

  • The pill turns amber while you give the instruction.
  • If it can't help, your text is left unchanged and the instruction is kept in History.

Where to change it

History

Saved dictations and activity, newest first. History is stored on this PC; speech and AI may still process content through your selected services.

  • Use Copy or Type it again where available. Advanced view adds entry actions such as Delete; a single-entry deletion offers Undo for five seconds. Read aloud requires ClaraVoca and a supported profile.
  • With the relevant spoken controls enabled, “Voca again” repeats the last dictation. Hands-free listening depends on your profile, settings and the focused window.

Where to change it

Tasks and notes

Press and say it: “remind me to call the dentist at ten”. Review the saved task or note and its interpreted date and time.

  • A due reminder can appear with LexVoca in the tray, provided the app is running and the PC is awake.
  • Snooze it, mark it done, or set it to repeat.

Where to change it

Trigger phrases

Say a short phrase, such as “email signature”, and LexVoca types a whole saved block of text.

  • Pick two or three words you wouldn't say by accident.
  • The saved text can be several paragraphs.

Where to change it

Opening programs by voice

Say “launch Word” or “switch to Outlook” and LexVoca opens it, or brings it to the front.

  • Only what you have installed: programs on your Start menu, and apps you add yourself.
  • Use a configured starting word, such as “launch”. Preview the phrase on the Launch page to check what would open without opening anything.
  • Two matches? LexVoca asks; say the number or click.

Where to change it

Where your text lands

With Type where I started on, LexVoca brings the original window forward before inserting. With it off, insertion uses the window focused when dictation finishes.

  • Keep the intended field selected. If the window closes or insertion is refused, look for the recognised text in History.
  • Clipboard-based insertion can restore your previous clipboard text when Put my clipboard back is on. A newer copy is left alone; failed insertion may leave the dictated text on the clipboard for recovery.

Where to change it

Hands-free: no keyboard at all

Enable hands-free listening in Settings, Typing & pasting. In a supported focused console or browser, say “Voca listen” to start; a pause ends that dictation.

  • “Voca again” types your last dictation; “Voca paste” pastes what you copied.
  • The phrase listener runs on this PC and uses the Windows default microphone. The selected speech service processes the dictation that follows; phrase recognition can still mishear speech.

Where to change it

Reading aloud with ClaraVoca

ClaraVoca, the NeuroDynamic reader, reads LexVoca's text aloud with highlighting.

  • Use Read aloud in History and in Tasks & notes.
  • ClaraVoca must be available on this PC. Reading controls are shown only in supported profiles and when the companion is available.

Where to change it

Where your speech is turned into text

Choose NeuroDynamic hosted, a provider or your own speech server, or speech on this PC. Each choice keeps its own speech settings.

  • NeuroDynamic hosted: sign in with your NeuroDynamic account.
  • Provider or own server: OpenAI, Groq or your own, with your own key.
  • On this PC: audio is recognised locally. Model and engine downloads require a separate request. The engine checks for usable GPU acceleration and falls back to CPU with an explanation when needed. Accuracy and speed vary; terminal input is unavailable in this mode.
  • AI is separate: tidying, Your Voices and voice edit use your NeuroDynamic account even with on-PC speech. Turning on local speech does not turn these into offline AI features.

Where to change it

Feedback and backup

Tell us what to improve from inside LexVoca, and keep a copy of your set-up.

  • Feedback: review the preview before sending. You can choose to include your last five dictations. Log upload is unavailable; use Settings, Backup & more, Export diagnostics to create a separate troubleshooting file.
  • Backup: save selected setup data to a file; account backup needs service support and access. Secret key fields, credential references, server destinations and Auto-enter are excluded. Review personal text in any phrases or History you choose to include.

Where to change it

Commands in terminals

In Technical, supported speech can become a command for review. Auto-enter is off by default; Bash always needs draft review and manual Enter.

  • Spell it (Bash): “L S dash L A” types ls -la.
  • Say the words (PowerShell): “get child item” types Get-ChildItem.
  • Just ask (Bash): “show disk space” types df -hT. Other shells can produce a different command; check the selected shell.
  • Built-in commands are checked against an allowed list. Saved phrases can change things; review them before running. AI suggestions need AI access and are labelled for review.
  • In other supported shells, Auto-enter can submit eligible basic checks and folder navigation after target checks. It pauses if an unsubmitted line is left behind. An allowed command is not a guarantee of safety.
  • In Bash, careful drafts show the exact command before anything is typed.

Where to change it

Virtual machines and remote consoles

Some windows, such as VM consoles and remote desktops, ignore pasting. LexVoca types into them key by key instead.

  • Superpaste () types what you copied into any of these windows.
  • When the VM grabs the keyboard, try a configured mouse button, the tray menu or Paste in 3 seconds. Whether a host shortcut or mouse button reaches LexVoca depends on the console.
  • Work inside VirtualBox watches for the configured Superpaste shortcut while a VirtualBox console is in front. This optional workaround is specific to that console; it is not a general VMware or remote-session hook.
  • Browser consoles (for example a hypervisor's web console) are web pages; LexVoca types into them with VM / browser keys.
  • Host and guest keyboard layouts must match for typed text to come out right.

Where to change it

Apps and websites that need a different method

Most apps need nothing. If words don't arrive properly in one window, teach LexVoca how to type there.

  1. Open Apps & websites and press Check in 3 seconds, then click into the window.
  2. LexVoca shows how it would type there. Pick another method if needed.
  3. For a web page, add a website rule; it wins over the browser's own setting.

Where to change it

Saved