Screenshot anything. Never leave the tab.
Press one shortcut. The page holds still, you drag a box around what matters, scribble on it, and it's on your clipboard, ready to paste.
Not a mockup. The extension's own code drew this, and a script captured it. See For developers.
Opera has a screenshot tool that gets out of your way. One shortcut, the page freezes, you drag a box, put a red arrow on the thing you mean, and paste it into a chat. You never leave the tab and you never think about it.
Most other browsers don't have that. The extensions that try tend to lose the part that mattered: they open a new tab, or make you save a file and open it somewhere else to draw on it, or want you to sign up first. Screenshots aren't hard. The problem is that a two-second habit turns into a twelve-second chore, and it breaks the thought you were having.
NhakoCapture brings the two-second version back. It's part of the Nhako family, alongside Nhako Tools and NhakoSearch: free, private, and made to feel good to use.
| The page holds still | The screen is captured before anything of ours appears, so the page stops exactly as it was, and our buttons can never end up in your screenshot. |
| Drag to frame | Draw a box, then fine-tune it with eight round handles, by dragging the middle, or with the arrow keys. A small label shows the real size of the image you'll get. |
| Mark it up right there | Pen, arrow, blur, highlighter and text, on top of the page you're already looking at. Seven colours, three line widths, and a magnifier for the fiddly bits. |
| Blur that really hides things | Blur replaces the pixels with a blurred copy of themselves. Nobody can un-blur the file you share. |
| Undo everything | Every mark can be taken back, one at a time, all the way to the start. |
| Copy or save | Straight to your clipboard as a real image, or to a proper Save dialog where you choose the folder. |
| The whole page, too | It scrolls the page for you and stitches it into one tall image. Sticky headers show up once, not on every screenful. |
| Or a PDF | Sends the entire page to your browser's print preview, with Save as PDF ready to go. |
| Private by design | No internet connection, no account, no analytics. Nothing leaves your computer. |
![]() |
![]() |
![]() |
![]() |
![]() |
|
Every product shot is drawn by the extension's own code in a real browser, not mocked up. See For developers for how they're made.
Add NhakoCapture from the Chrome Web Store. It's free, and it works in Chrome, Brave, Edge and other Chromium browsers. Click Add to Chrome (Brave and Edge say the same, or Get), then pin it from the puzzle-piece menu so it's always one click away.
If you'd rather run it straight from this repository, load it yourself. It takes a minute:
- Download or clone this repository.
- Open your browser's extensions page:
chrome://extensions,brave://extensionsoredge://extensions. - Turn on Developer mode (top right).
- Click Load unpacked and pick the folder that contains
manifest.json. - Pin NhakoCapture from the puzzle-piece menu so it's always one click away.
Needs Chrome, Brave or Edge 116 or newer.
Press the shortcut, or click the camera in your toolbar.
| Shortcut | |
|---|---|
| Windows, Linux, ChromeOS | Ctrl Shift 5 |
| Mac | ⌘ ⇧ 2 (macOS keeps ⌘⇧5 for its own screenshot tool) |
Your browser might pick a different shortcut if another extension already uses
that one. Hover over the toolbar icon to see which one you have, and change it at
chrome://extensions/shortcuts.
The page dims and holds still. Then:
- Drag over anything to frame it.
- Visible area frames everything on screen.
- Whole page scrolls the page and captures all of it, then opens it in a new window for you to mark up.
- Save as PDF opens the print preview for the entire page, with Save as PDF already chosen.
Once you've framed something, the tool bar slides up underneath. Pick a tool, draw, then hit Copy or Save. You'll see a quick flash and a tick, and the overlay gets out of your way. You can keep adjusting the frame's edges while a tool is selected. A stray click outside the frame won't throw your drawing away.
Press the shortcut again at any time to close it.
| Key | What it does |
|---|---|
Ctrl Shift 5 / ⌘ ⇧ 2 |
Open NhakoCapture, or close it again |
P A B H T Z |
Pen, Arrow, Blur, Highlighter, Text, Magnifier |
Ctrl Z / Ctrl Shift Z |
Undo / redo (⌘ on a Mac) |
Ctrl C |
Copy and close |
Ctrl S |
Save and close |
← ↑ → ↓ |
Nudge the frame by 1 pixel |
Shift + arrows |
Nudge by 10 pixels |
Alt + arrows |
Resize the frame |
Esc |
One step back: put the tool down, then clear the frame, then close |
They're the slower option, and here's why:
- It takes about half a second per screenful. Browsers only allow about two
captures a second. The toolbar icon counts along while it works:
3/9, or a percentage on long pages. - It scrolls through the page first, so pictures that only load when you scroll to them are there in the final image.
- Sticky headers, cookie banners and chat bubbles appear once, at the top, instead of repeating down the whole image.
- Switching tabs pauses it. It carries on when you come back, so it never captures the wrong tab by mistake.
Escstops it. The page goes back exactly where it was, whether you stop it, it fails, or it finishes.- Very long pages are cut off at 16,384 pixels instead of being shrunk into a blurry mess, and the editor tells you when that's happened.
- If the whole page already fits on screen, the button says so rather than giving you the same picture twice.
| Permission | Why it's needed |
|---|---|
activeTab |
To capture the tab you're on. Only granted when you press the shortcut or click the icon, and only for that tab. |
scripting |
To put the overlay on the page you're capturing. |
storage |
To hand a capture to the editing window on pages where the overlay can't run. |
downloads |
To open a real Save dialog, so you pick where the file goes. |
clipboardWrite |
Firefox only, in place of offscreen: lets the add-on's background page put the image on the clipboard. |
offscreen |
Chrome only. To copy images on plain http:// pages, where the page itself isn't allowed to use the clipboard. On those pages the copy pastes into chats and documents, but not into image editors, and NhakoCapture tells you so. |
No host permissions, no access to your browsing history, and no network requests. See PRIVACY.md.
- Coming to Firefox. The same code now builds for Firefox too
(
./tools/package.sh firefox). Firefox has no offscreen documents, so there its background page does the copying and saving itself. - The camera mascot in the toolbar. The small mark at the start of the toolbar is now the pixel camera from the icon, instead of the old butterfly.
- Fixed: opening two editing windows close together could show the wrong capture, label a normal capture "cut off", or keep a whole-page image in memory until the browser restarted.
- Fixed: pressing the shortcut twice very fast said the page was off-limits. It now says to try again.
- Fixed: a page too big to encode opened an empty editing window. The toolbar now says the page is too big.
- Tidier code. Unused exports and design tokens are gone, and the toolbar
builds its icons as DOM nodes, not through
innerHTML.
- Ready for the Chrome Web Store. A privacy policy, a full store listing, and screenshots and promo tiles at the store's exact sizes.
- A solid Copy button. The Copy button and the tick after a copy are now one flat brand pink instead of a pink-to-lavender gradient. The store graphics and the launch poster switched to solid pink to match.
- Fewer permissions. The
debuggerpermission is gone. Save as PDF now opens your browser's print preview instead, so installing NhakoCapture no longer warns that it can "read and change all your data on all websites". - Fixed: a capture could disappear if the overlay failed to start on an unusual page. It now opens in the editing window instead.
- Tidier code. Old comments and a message nothing used are gone, and the
version number now lives only in
manifest.json.
- A new look. A frosted, dusk-tinted toolbar in the Nhako pink and lavender, round handles, and small spring animations when things appear. A shutter flash and a tick tell you the copy worked. Everything still respects your system's reduce motion and high-contrast settings.
- A new icon. A butterfly caught in a capture frame, drawn to stay sharp even at 16 pixels.
- Plainer words. "Visible area", "Whole page", "Pen", "Highlighter", "Magnifier". Messages say what happened ("Couldn't copy that. Try again?") instead of "Copy failed".
- Fixed: switching tabs during a whole-page capture could stitch the other tab into your image. It now pauses until you come back.
- Fixed: some websites' own styles could shrink or move the overlay, leaving it out of line with the page. The overlay now holds its position whatever the site does.
- Fixed: after Visible area, the top bar couldn't be clicked. It now just fades back and comes forward when you point at it.
- Fixed: the tooltip always said
Ctrl+Shift+5, even on a Mac where the shortcut is⌘⇧2. It now shows the shortcut you actually have. - Fixed: cancelling the Save dialog could show an error on some browser versions. Changing your mind isn't an error.
- Faster long pen strokes. Drawing a long line no longer slows down as it grows.
- Screenshots are now named like macOS names them:
NhakoCapture 2026-09-29 at 14.03.22.png.
No bundler, no framework, no dependencies. The code you read is the code that runs.
manifest.json Manifest V3
PRIVACY.md the privacy policy the Web Store listing links to
icons/ made from docs/brand by tools/make-icons.py
docs/brand/ the logo (SVG masters and a PNG)
docs/launch/ the launch poster, as HTML and PNG
docs/store/ Chrome Web Store listing: LISTING.md, the graphics,
their HTML sources (src/) and product shots (raw/)
src/
├── background.js captures FIRST, then adds the overlay; routes messages
├── offscreen.js clipboard fallback and blob URLs
├── lib/
│ ├── namespace.js module registry (injected scripts can't use ES modules)
│ ├── geometry.js CSS pixel ⇄ device pixel maths
│ └── clipboard.js writes the PNG from the page you clicked in
├── engine/
│ ├── ops.js the list of marks; undo and redo
│ ├── render.js replays the marks over the image
│ └── stitch.js joins whole-page tiles into one image
├── overlay/
│ ├── inject.js start-up, keyboard, tidy-up
│ ├── stage.js shadow root, frozen backdrop, dimming, flash
│ ├── selection.js frame, handles, size label
│ ├── fullpage.js the scroll-and-capture loop
│ ├── annotate.js drawing tools
│ ├── rail.js tool bar
│ ├── toolbar.js top bar
│ ├── tokens.css design tokens ("dusk glass")
│ └── overlay.css overlay styles and motion
└── fallback/ the editing window, for browser pages and whole-page captures
tools/ tests, packaging, store graphics, icon and poster rendering
Three decisions shape everything else:
Capture first, then draw. The screen is captured before the overlay exists. That's why our own buttons can never appear in your screenshot, and it's what makes the page "freeze".
Everything lives in a shadow root. No website's styles can reach the overlay, and the overlay's styles can't disturb the website. The one thing a shadow root doesn't shield, the host element's own size and position, is pinned inline so a site's CSS can't move it either.
Marks are records, not paint. The final image is made by replaying a list of marks over the capture. That's what makes undo possible.
The look comes from the Nhako family. Nhako Tools gave
it the brand pink (#FF91E7) and the rule of one typeface, San Francisco,
through the system and never downloaded. NhakoSearch and NeraOS gave it lavender
(#C3A6F0), the plum ink, and the striped progress bar. The mascot, a purple
pixel-art camera, is one of the seven Nhako mascots. The
one button that finishes the job, Copy, is solid brand pink with plum
text (8.75:1), so there's never any doubt where to click. The feel
comes from macOS's own screenshot tool: frosted dark glass, hairline edges, and
springs that settle instead of bouncing.
The toolbar stays dark on purpose. It has to be readable over a white article and
over a black video. Every text colour's contrast is worked out and written next
to it in tokens.css, and each one clears WCAG AA even
with a pure white page behind the glass.
./tools/test.shThat runs everything: about 475 unit checks and 218 checks in a real browser.
node tools/preview.mjs out/This drives a real headless Chromium with real mouse and keyboard input, using nothing but Node 22. It checks things a unit test can't: that the dimming leaves the framed area pixel-for-pixel untouched, that blur really destroys detail, that a website's own CSS can't move the overlay, and that pressing the shortcut three times leaves exactly one overlay on the page.
It also makes the screenshot at the top of this page:
node tools/preview.mjs out/ --heroBuild with ./tools/package.sh firefox and follow
docs/store/FIREFOX.md. ./tools/package.sh all
builds both zips from one version number.
-
Bump
"version"inmanifest.json. The store refuses a version it has already seen. -
Build the upload. This runs every check first, then zips only
manifest.json,icons/andsrc/:./tools/package.sh
-
Upload
dist/NhakoCapture-<version>.zipin the developer dashboard. For the first release, fill in every field fromdocs/store/LISTING.md.
To rebuild the brand assets:
python3 tools/make-icons.pynode tools/render-poster.mjsTo rebuild the Chrome Web Store graphics. The first script drives the real overlay in headless Chrome, including a real scroll-and-stitch, to capture the product shots. The second frames them into the screenshots and promo tiles at the store's exact sizes, as 24-bit PNGs with no transparency:
node tools/store-shots.mjs && node tools/render-store.mjsOn purpose, and not by much:
- Zoom is a magnifier, not a zoomed canvas. The overlay sits exactly on top of the live page, so zooming it would pull the frame away from what it's showing. The magnifier follows your pointer; click to park it, click again to let it go.
- No stickers or emoji. A lot of extra weight for very little.
- No selfie camera. It would need webcam access, and the whole promise is that nothing leaves your computer.
- Whole-page captures go down, not sideways. Wide pages are captured at the width of your window, and scrolling boxes inside a page are captured as they look on screen.
- Endless feeds aren't chased. A page that keeps growing as you scroll is captured as far as it existed when you started.
Version 2.2, live on the Chrome Web Store. Everything described above works. The browser checks drive a real Chromium with real input, but they don't replace trying the extension in your own browser, so do that hands-on pass with the packaged zip before each release.
Made by kimzam, to bring the Opera feel to whichever browser you like. Part of the Nhako family.






