Navisual User Manual

Quick answers

Does Navisual click or type for me?

No. Navisual only shows you what to do: a pointer on the control, a caption and a spoken instruction. You make every click and keystroke yourself, so you can stop, change course or ignore it at any point.

Is Navisual free?

It is free to download and includes 30 free requests per computer, with no sign-up and no API key. A request is one call to the AI, and one call can cover more than one step. After that you can buy coins, use your own API key, or run a local AI model for free.

Do I need an account?

No. The free requests, your own API key and a local model all work without one. You need an account to buy coins, and it keeps your coins and purchases with you across computers.

How much does Navisual cost after the free requests?

Coins start at $5 for 1,000 coins. A request costs 6, 12 or 18 coins depending on the quality tier, so $5 buys about 55 to 165 requests. Coins do not expire. With your own API key the provider bills you directly, and a local model costs nothing.

Which apps does Navisual work with?

Navisual works with almost any Windows app. It has been tested in Microsoft Word, Excel, PowerPoint, Outlook and OneNote; Chrome and Edge; File Explorer and Windows Settings; and Blender, Krita, VS Code, VLC and LibreOffice. In apps that paint their own interface, such as Photoshop and SolidWorks, the AI points from what it sees in the screenshot.

Which versions of Windows does Navisual support?

Windows 10 (version 1809 or later) and Windows 11. macOS and Linux are not supported yet.

Is Navisual watching my screen all the time?

No. Navisual takes a screenshot for the AI when you ask for help or ask for the next step. The exception is Autopilot, which you turn on yourself: it compares the screen on your computer about twice a second, and goes to the AI only when the screen has changed enough to look like you finished the step.

What does Navisual send to the AI?

A screenshot of the window you chose, with everything else greyed out, plus what you typed, the window title, the names and states of the controls in that window, the control you last clicked there, and your Windows version. The Privacy Policy has the full list.

Can Navisual run offline?

Yes, with a local AI model through Ollama or another local server such as LM Studio. Nothing leaves your network and nothing is logged. You need a model that can read images, and the quality of the guidance depends on that model.

Which languages does Navisual support?

The interface is in English, Simplified Chinese and Traditional Chinese. The AI answers in the language you type in, and reads the answer aloud if Windows has a voice installed for that language.

1. Getting started

What you need: Windows 10 (version 1809 or later) or Windows 11, and an internet connection unless you run a local model. Navisual also needs WebView2, which ships with Windows 11 and installs automatically on Windows 10.

Install. Download and run the installer from the home page. It’s code-signed, but a newly-released app still trips Microsoft SmartScreen. Click More info → Run anyway if you see it.

First launch. You’ll see a short privacy notice explaining what Navisual captures, where it goes and what is logged. You don’t need an account: the free tier signs you in anonymously and gives you 30 requests to try it with.

Then the loop is always the same:

1. Open the app you want help with and leave it visible.
2. Type what you want to do in Navisual’s box, in plain language, e.g. “make a pivot table from this data” or “add a bevel to this cube”, and press Guide me. The box may already hold a suggested task: use it or type over it.
3. Follow the pointer. Navisual takes a screenshot, works out the next step, and draws a marker on the exact button, with the instruction spoken and captioned.
4. Do the step yourself, then press Ctrl+` for the next one.

Navisual only captures a screenshot for the AI when it needs one: when you ask a question, or when you ask for the next step. The one exception is Autopilot, which you switch on yourself.

When the task is done, press ＋ New task to start clean. To change or correct what you asked for, just type in the box and press ↩ Follow up.

Where to put the panel

Navisual works wherever the panel is. These are suggestions, not requirements: what works best is a panel that doesn’t cover the app you are working in.

• One screen: docking or collapsing works best. Dock left or Dock right, in the ··· menu, gives Navisual a strip down one side and the rest of the screen to your app, which suits a wide screen especially well. Or collapse Navisual to its goldfish icon and press Ctrl+` for each next step: the pointer, caption and speech carry on as before. On a smaller screen, collapsing may suit you better.
• Two or more screens: we recommend putting the panel on a screen of its own, with the app you are working in on another. Your app keeps its whole screen, and the steps stay in view beside it.

2. The panel

The Navisual panel is the window that holds your task and the steps so far.

Dock left and Dock right, in the ··· menu, put Navisual down a quarter of the screen at full height and give the rest to one app you choose, so the panel never sits on top of your work. Undock floats it again.

3. Choosing what Navisual sees

By default Navisual looks at whichever window is in front. The app chip in the title bar shows which app that currently is. A brief outline marks the window when you pin it, so you can see what’s being shared.

Click the app chip to change it. You get three kinds of choice:

• A specific window. Picking one pins it (📌): Navisual keeps looking at that window even when you click elsewhere. This is the right choice whenever you’ll be switching between windows, or when the app opens dialogs that steal focus. A minimized window is listed too, and picking it brings it back.
• Entire desktop (or Screen 1 / Screen 2 on a multi-monitor setup), for tasks that span apps, like dragging a file from Explorer into another program. On multiple monitors, pick the individual screen rather than the whole desktop: stitching every screen into one image shrinks the detail past the point where the AI can read it.
• Auto-detect to go back to following the foreground window.

When one window is shared, everything outside it (including other apps overlapping it and Navisual’s own panel) is greyed out before the image is sent, so the picture shows your target app and nothing else. What else goes with it is in section 14.

If a pinned window closes or is hidden, a banner says so with a Pick a window button. Navisual won’t look at anything in its place until you choose. If the window comes back, the banner clears itself. A pinned window that is only minimized needs restoring from the taskbar.

If nothing is open to guide, Navisual offers Guide me on this screen or Pick an app. It won’t widen what it looks at unless you choose that.

4. Following the guidance

Each step gives you an instruction, spoken aloud with a caption on screen, and, when Navisual can pin down the exact control, a marker on the button itself.

If no marker appears, you’ll see a quiet note saying the pointer isn’t available. That’s deliberate. Navisual would rather show you nothing than point at the wrong thing, and some surfaces genuinely can’t be read from the outside: a 3D viewport, a drawing canvas, a game. The written instruction is still correct; only the arrow is missing.

→ Next moves to the following step. Autopilot (in the action row) does the same automatically whenever the screen changes enough to look like you finished the step. It’s off every time the app starts, on purpose: it can cost a request each time it fires. Autopilot suits long, obvious sequences; manual stepping suits anything where you want to think between steps. If it fires too often or not enough, move the slider in Settings → Screen Guide → Autopilot.

Autopilot pauses itself in two cases: when the screen keeps changing without a clear next step, and once when you’re down to 5 free requests, so it doesn’t spend the last of them on its own. The panel says why, and turning it back on is up to you. It stops when your free requests run out.

Anything you’d have to type is put on your clipboard. A terminal command, a URL, a filename, text for a form field. When a step involves typing, the exact text is copied for you and the step shows a 📋 badge, so you paste rather than retype it. Keyboard shortcuts are the deliberate exception: those are named in the instruction, because a shortcut has to be pressed, not pasted. Navisual never runs a command or presses anything for you: you read it, then paste it and press Enter yourself.

Tasks that span several apps work without any setup: download something in the browser, edit a file, run a command in a terminal, and Navisual follows whichever window you’re working in. If you’d rather it stayed fixed on one app while you move around, pin that window (section 3).

Sometimes Navisual asks you a question instead of giving a step. Type the answer in the reply box and press Enter. If it can’t see enough to be sure, the honest answer, “bring the right window forward”, is what you’ll get.

What the panel’s messages mean

5. Hotkeys

Hotkeys are the smooth way to use Navisual, because they never move keyboard focus away from your app. Clicking Navisual’s panel does: Windows then eats your next click on the target app, because that click is spent re-activating the window, and a menu you had open closes. If you’ve ever had to click a button twice, that’s why. Use the hotkey and it doesn’t happen.

` is the backtick key, left of 1 and below Esc. The last two ship unassigned deliberately, so they can’t clash with a shortcut you already use. Set your own in Settings → Hotkeys, where you can also clear any binding with the ✕.

Next and Mark wrong are only held while a step is being guided, so with no task running those keys still belong to your other apps. If a step tells you to press a shortcut that is also a Navisual hotkey, Navisual lets go of it for that step and says so in the panel.

6. When it points at the wrong thing

Under each instruction there’s a ✗ This is wrong button. Pressing it asks what went wrong, and the answers do genuinely different things: picking the accurate one gets you a better fix, faster:

None of these fit? Type what’s wrong in the box and press ↩ Follow up. A typed note always goes to the AI. If Navisual is looking at the wrong app, click the correct window first, then press ✗. Your choice is logged so we can improve the guidance, never your screen; what you typed is logged with it unless you switch that off (section 14).

Several numbered boxes instead of one arrow? That means Navisual found more than one good match (three cells that all say “Total”, say) and it isn’t going to guess. ① is its best guess. Click the one you meant: your click is the answer, and there’s no dialog to dismiss. If it’s none of them, press ✗.

7. Recent tasks and exports

Navisual saves each task on your computer as you go. 🕓 Recent tasks lists the 20 most recent: click one to reopen it and carry on from where you stopped. Older sessions go to the Recycle Bin with their screenshots.

A saved session holds the conversation, your task, the plan, what you clicked, and each step’s screenshot. To stop keeping the screenshots, turn off Settings → Screen Guide → Saved sessions; a reopened session then re-reads the screen at its next step. Either way, a reopened session follows the app you click into next.

Export in Recent tasks saves your sessions, or just the ones you select, as one self-contained HTML file each, readable in any browser. Open a file… shows such a file in the panel; it joins your recent tasks only if you carry on working in it. Screenshots are pictures of your screen, so look through a file before sending it to anyone.

8. Language, speech and voice

Interface language. Navisual’s interface is in English, Simplified Chinese and Traditional Chinese. It follows your Windows language; change it in Settings → Screen Guide → Interface language.

The AI’s language. The AI answers in the language you type in, whatever the interface is set to.

Spoken instructions use the voices installed in Windows. On Auto, each reply is spoken in its own language with an installed voice for that language. Choose a preferred voice, or turn speech off, in Settings → Audio. To mute it for now, or hide the caption, use the ··· menu.

Voice input lets you speak your request: press Ctrl+D or the 🎤 button. It needs a microphone and an internet connection, because your audio goes to Microsoft’s online speech service. You can turn it off in Settings → Audio. Speech recognition can’t detect a language before you speak, so if yours isn’t your Windows language, set Voice language there. The choices include English, French, German, Spanish, Japanese, Chinese, Korean and Portuguese.

9. AI providers and cost

Navisual works with several AI backends. Choose in Settings → Provider.

• Managed: free. No key, no account. 30 requests per device. Note: free-of-charge AI models generally retain requests (including your screenshot) to train on. If that matters to you, use any option below instead. (Details in the privacy policy.)
• Managed: pay as you go. Top up from $5 and Navisual routes to strong paid models for you, with no API key to manage. Paid providers state they don’t train on your data.
• Bring your own key. Anthropic, Google Gemini, OpenAI, DeepSeek, Qwen, or any OpenAI-compatible endpoint. Requests go straight from your machine to that provider, not through our servers, and they bill you directly. Two notes: DeepSeek sees your screen only with deepseek-v4-flash-vision-exp, the model marked “sees the screen” in Settings. With the other DeepSeek models Navisual sends no screenshot, so the AI works from your words, the window title and the names of the controls, without the picture. And Qwen is the recommended cloud option for mainland China.
• Ollama or another local server. Runs on your own machine or network. Nothing leaves it, nothing is billed, nothing is logged, and it works offline. You need a vision model such as llama3.2-vision, or any OpenAI-compatible local server (LM Studio, llama.cpp, vLLM) through the Custom provider. Quality depends entirely on the model you run.

Coins and quality tiers

On the managed tiers you buy coins. A coin is $0.005, fixed. Your quality tier decides which model answers and what a request costs:

So $5 buys 1,000 coins, roughly 55 to 165 requests. Buy coins in Settings → Account ($5 to $500; checkout opens in your browser, and you need an account). Coins do not expire and are non-refundable except where the law requires it; see the Terms. When a signup offer is running, the app shows it.

Free requests are always used first, whichever tier is selected. When you’re down to 5, the panel says so.

Which to pick?

If you’re trying Navisual out, the free tier is fine. If you’re using it on real work, pay-as-you-go or your own key: both are markedly more accurate, and both keep your screen out of training data. If your screen content can’t leave your machine at all, a local model is the only option that guarantees it by construction.

Your balance is in the title bar; click it to reach billing. The Usage tab under ⓘ About shows your balance, and estimates what your own keys have spent.

10. App-specific support

Navisual reads most Windows apps through the accessibility information they already publish, which is why buttons, menus and ribbons work well nearly everywhere. Some apps need more.

Browsers are one of the best cases: web pages publish rich structure, so pointing inside a web app, form or admin panel is usually exact. Chrome and Edge get a dedicated fast path.

Terminals (Windows Terminal, PowerShell, a shell inside your editor) work through the clipboard rather than through pointing: there are no buttons to aim at, so Navisual gives you the command to paste instead (section 4). It never runs one for you.

Excel, Word and PowerPoint are handled specially: Navisual can resolve a cell reference, a phrase in your document, or a shape on a slide to exact pixels, rather than guessing visually. Nothing to set up.

Blender offers an optional add-on. Blender draws its whole interface itself, so from the outside it’s just an image: the add-on lets Blender tell Navisual where its tools are, and which mode and brush you’re in. It makes tool-shelf and properties-tab guidance far more reliable.

• Navisual offers to install it when it notices you’re working in Blender. Click Install, and it copies the file into Blender’s add-ons folder.
• Then restart Blender, open Edit → Preferences → Add-ons, search for Navisual, and tick its checkbox. That tick is the consent step; Navisual can’t and won’t do it for you.
• It is read-only: it answers questions about the interface and cannot click, type, run commands, or touch your file. It listens only on your own computer and sends nothing to the internet.
• Each Blender version has its own add-ons folder, so install it once per version you use. Untick it any time to turn it off.

Apps that paint their own interface (SolidWorks’ command ribbon, the Photoshop and GIMP tool boxes, 3D viewports, most games) publish nothing for Navisual to read locally. This does not mean Navisual can’t guide you in them. When there’s nothing to read, the AI falls back on looking at the screenshot the way you do, and points using what it sees. On a capable model that works well, including on interfaces it has never encountered.

What changes is how it points: the highlight tends to be a region around the control rather than a tight outline of it. And it depends much more on the model than anywhere else in the app: reading an unfamiliar interface visually is a genuinely harder job than reading a label. If an app like this isn’t giving you pointers, raise the quality tier in Settings → Provider (or use your own key with a current vision-capable model). That single change matters more here than any other setting.

Note that the parts of these apps you most often need help with (menus, preferences, dialogs, property panels) are usually not self-painted, and behave normally. It is common for the same app to be exact in its menus and approximate on its canvas.

Very small text is the other hard case: on-screen text recognition has a size floor, so tightly-packed interface labels can be missed. You still get the written instruction either way.

Learned notes. When finding a command in an app takes more than one try, Navisual writes down where it was, so the guidance gets it right the first time next time. The notes hold interface labels only, stay on your computer and go with your later requests in that app. Read, change or delete them in Settings → Screen Guide → Learned notes.

11. Settings

Open Settings with the ⚙ button in the title bar. There are five tabs:

Apply saves and keeps the window open; OK saves and closes. No restart is needed. Reset all settings restores every tab, not just the one you’re on: your API keys are kept, but server addresses and model choices are not.

12. Updates, your data, uninstalling

Updates. Navisual checks for a new version shortly after it starts, and shows a notice in the title bar when there is one. You can also check in ⓘ About → Check for updates; then press Install & restart.

Your data. Settings, saved sessions, learned notes and logs are in %LOCALAPPDATA%\com.navisual.app. API keys and your sign-in are kept separately in Windows Credential Manager, encrypted for your Windows user.

Uninstalling. Remove Navisual from Installed apps in Windows Settings (Apps & features on Windows 10). If your saved sessions and settings are still there afterwards, delete the folder above.

Deleting your account. Go to Settings → Account → Delete account. This is permanent, and coins are not refunded. What happens to stored records is in the Privacy Policy.

13. Troubleshooting

14. Privacy in one screen

• Screenshots go to the AI only when you ask a question or ask for the next step, or when Autopilot, which you turn on yourself, sees the screen change. Autopilot’s checks for a change stay on your computer.
• When one window is shared, everything else on screen is greyed out before sending. If you choose to share a whole screen, that whole screen is sent.
• Along with the picture go what you typed, window titles, the names and states of the controls in that window, the control you last clicked there (its name, never its contents) and your Windows version. Navisual does not monitor your keyboard.
• Your screenshot goes to the AI provider you chose. On the managed tiers it passes through our relay. With your own key it goes straight to that provider, and with a local model it stays on your machine or network.
• We don’t store your screenshots on our servers. Each step’s screenshot is kept on your computer with your saved sessions (you can switch that off). Settings stay on your machine, and API keys in Windows Credential Manager.
• We log the task you type and how each step went, never a screenshot. You can switch off the typed task in Settings → Screen Guide → What you type; with an AI on your own machine or network nothing is logged.
• Voice input, if you use it, sends your audio to Microsoft’s speech service.
• The free tier’s models may train on what you send. Paid, BYOK and local options don’t.

The full details, including the Blender add-on, are in the Privacy Policy.