Screenshots and OCR for Windows
A screenshot tool
that can read.
Drag a region and KeptInFlow gives you the picture or the words on it, on this machine, with no upload. Annotate it, capture the whole scrolling page, and find it again months later by typing what it said. Pay once, keep forever.
If Windows SmartScreen asks about this new release, click More info → Run anyway. The installer is code-signed and verified.
Text recognition runs on this PC. Nothing is uploaded to read it.
One install replaces
- Snipping Tool
- OCR utility
- Annotation editor
- Screen recorder
- Screenshot folder
- Color picker
Runs on your PC, not in a tab · no account · pay once
The moment that sells it
Copy text out of anything
you can see.
An error dialog. A PDF that will not select. A screenshot somebody pasted into Slack. A video frame. Drag over it, press Copy text, and the words are on your clipboard. Recognition happens on this machine, so it works with no connection and nothing is uploaded to do it.
Drag a region, anywhere
The overlay freezes the screen so nothing moves under you, then you drag with precise handles, a magnifier for the pixel, and live dimensions. Tap a window instead of dragging to take exactly that window.
Or snip with just the mouse
Hold Ctrl, drag with the right button, and the selection starts where your hand already was. The modifier and the button are both yours to change, and the gesture is off until you switch it on.
On-device text recognition
Text recognition runs on your PC. No folder to configure, no cloud call, no per-scan cost.
AI cleanup when the layout is messy
Copied text is tidied by a cheap, fast model: wrapped lines rejoined, swapped characters fixed, spacing evened out. It sends only the words it read, never the picture, and one switch in Settings, Screen capture turns it off. Plain OCR never needs it.
Or skip the region and grab the words
Hold Alt and drag across words
Anywhere on screen, with nothing to open first. Let go and the words are on your clipboard. Alt with a plain click still goes to the app underneath. On from the start; change the key or turn it off in Settings, Screen capture.
The whole screen, as words
One key freezes the screen and turns all of it into text you can drag over. Double-click takes a word, triple-click takes the line, and Ctrl+C copies.
Pick words inside a snip
After you drag out a region, press Select text and the frozen picture behaves like a page: select just the sentence you want. Copy text still takes the whole region when that is what you need.
Selects the way a page does
Drag down one column of a two-column PDF and you get that column, not the words across the gutter. The selection stays inside the window you started in, and it shows as one band per line.
When you let go, the toolbar offers Copy text, Copy image, Annotate, and Pin to screen, with Save to file and Share as a link under More, along with AI actions worked out from what is actually in the shot (an error, a table, code, another language). Every snip also goes on the Windows clipboard by default, and you can tell a snip to skip the toolbar entirely and copy the image, or the text, the instant you release.
Mark it up without leaving
Arrows, boxes, steps, blur.
The same drawing tools live on the capture overlay and in the full Annotate editor, so a quick arrow costs you nothing and a proper walkthrough is the same set of tools, larger. Marks stay editable: move them, restyle them, undo them one at a time.
- Arrow
- Box
- Ellipse
- Line
- Pen
- Highlight
- Text
- Numbered steps
- Blur
- Spotlight
- Crop
Every tool has a single-letter shortcut, and the letters are the same in both places.
Numbered steps
Drop numbered markers in the order you want the reader to look at things, for a bug report, a how-to, or a design note. Like every other mark they stay editable: drag them, restyle them, undo them one at a time.
Redact, and auto-redact
Blur anything sensitive by hand, or use the one-click pass that finds emails, card numbers, phone numbers, and keys in the capture on this PC and blurs them for you. Redactions are burned in before the image is read or shared.
Spotlight, crop, and frame
Dim everything except the part that matters, or cut the frame down to it. Rotate a crooked capture straight, and drop the finished image onto a background with a shadow when it is going into a doc or a post.
Copy the marks, not just the picture
Select marks and Ctrl+C copies those marks. Ctrl+V drops them onto the next image you open, scaled to fit it, so a set of callouts you drew once can be reused. With nothing selected, Ctrl+C copies the picture.
Marked on the overlay, finished in the editor
Draw a fast arrow while the capture toolbar is still up, then open the full editor if it needs more. The marks you already drew come with it, still editable.
Past the edge of the screen
Capture the whole page,
not the visible part.
Snip something that carries on past the bottom of the screen and the capture toolbar offers a Whole page chip. It only appears when there is actually more page to get.
Down, across, or both
A long article stitches into one tall image. A wide spreadsheet or Gantt chart stitches sideways with Across, and Both ways unrolls the entire scrolling area band by band.
Sticky headers drawn once
A header that stays pinned while the page scrolls is lifted out and drawn a single time instead of repeating down the image. A frozen left-hand column gets the same treatment.
Check the seams before you keep it
The stitched image opens right there, scrollable at full width so you can inspect the joins, with a Fit toggle that drops the whole page into view. Copy, save, share, annotate, or pin it from there.
Multiple monitors, mixed scaling
Snips and recordings land where you drew them even when one screen runs at 150% and the other at 100%. KeptInFlow captures the desktop in the same coordinate space Windows reports the cursor in, so an odd-scaled second monitor does not shift the result.
Color picker too
Alt+Shift+C freezes the screen under an eyedropper loupe. Click to
copy the pixel's value, and Shift swaps between hex and
rgb(…).
Scrolling capture ships without a global shortcut on purpose. Use Scroll on the capture toolbar, or bind a key for it under Settings, Screen capture.
The folder problem, solved
Find the screenshot
by what it said.
Every capture is kept and indexed by its text, so months later you type the words that were on screen instead of scrolling through a folder of dated PNGs. The search runs on this PC.
Search by content, not filename
Type what the picture said. The matching captures come back with the line they matched on.
A list that opens instantly
Hundreds of captures, and the window still opens straight away.
Open a search result and the matching words are
highlighted inside the picture, not just handed to you as a file. An
archive of hundreds opens as fast as an archive of ten. Your snips are
also reachable from the command palette, scoped with
:snip.
The obvious question
Why not Snagit?
Snagit is a good capture tool. The difference is what happens after the capture, and what happens when you stop paying.
| Snagit | KeptInFlow | |
|---|---|---|
| What it costs | $39/yr, with no perpetual option | One payment. The app is yours |
| If you stop paying | You stop having it | You keep the last version you got |
| Text off a capture | Grab Text, in the editor | On the capture toolbar, on this PC, straight to the clipboard |
| Finding an old screenshot | By file, folder, and date | By the words that were on it |
| Asking a question about it | Not what a capture tool does | AI actions offered from what is in the shot, with your own key |
| Everything else on your PC | Capture and recording | Meetings, dictation, clipboard, text tools, in the same install |
Price checked September 2026. Check it yourself; it moves.
Same purchase
The capture tool is one part of it.
KeptInFlow is one install with one price, and screen capture is a quarter of it. Here is the rest, briefly.
Meeting recording
Records Teams, Zoom, Meet, and Slack huddles from your own PC with no bot in the call, tells the speakers apart, and writes the summary.
Screen recording
Record a region to GIF or H.264 MP4, then speed up the dull parts, cut dead air, blur secrets, and zoom in, without a separate editor.
Dictation, offline
Hold a key, speak, release, and cleaned-up text pastes at your cursor. With the free on-device model it is unlimited and offline.
AI in every text box
Type a question between << >> in any app
and the answer replaces it. Edit selected text by describing the
change.
Clipboard cycler
Alt+Tab for your clipboard. Hold the modifier, flip through recent clips, release to paste the one you meant.
Command palette
One box that launches apps, switches windows, searches your snips, notes and meetings, or just asks AI.
See the whole app on the home page, or read what it watches for the feature-by-feature list of anything that reads your screen, mic, or clipboard.
No subscription required
Keep it free. Or pay once for Pro.
Try everything free for 30 days. Then keep the free plan forever, or go Pro.
App forever. The first 100 buyers get every update for life; after that, a license includes 1 year of updates. Keep the last version you got if you don't renew. This is not a stealth subscription.
Free trial
30 days
- Every feature, no limits
- Bring your own API key
- Then keep the free plan forever, or go Pro
- No card up front. Nothing to cancel.
Free
$0 forever
- Snip, grab text, scrolling capture, color picker
- Clipboard cycler, clean paste, text expansions, calculator
- Smart paste, verification codes, AutoCorrect, mouse gestures
- AI chats, rewrite, inline prompt, command palette
- Record your screen, with sound
- Dictation: 2,000 words a week
- Pair your phone and send things across
100 AI actions a month on us. Want more? Use the free on-device model, your own AI key, or the AI plan.
Pro
$59 one-time
$59 for the first 100 buyers, then $79, then $99.
- Everything in Free, plus Screen Memory, the journal, meetings and billing
- The Studio look, webcam and step-by-step recordings
- Unlimited on-device dictation, no quota to run out of
- Own it forever, no subscription
- Capture, OCR, annotation, and recording all included
- Every future update, for life
- 30-day refund, no questions asked
$59 is the founding price for the first 100 buyers, with lifetime updates. It ends when those seats sell out, not on a timer. The next 200 buyers pay $79, and after that the price is $99. Those two include one year of updates, renewable for $39/yr (or don't renew and keep what you have). Or get lifetime updates for $199. Every option is a single payment with the 30-day refund.
AI Included
$12 a month
AI that just works. Nothing to set up.
- 5,000 fast AI actions a month
- 1,200 minutes of cloud dictation a month
- Every Pro feature included while it runs
- No key to get and no AI account to open
- Works on its own, or on top of a Pro license
It renews automatically at that price until you cancel. Cancel anytime from Manage subscription in the app, or at app.lemonsqueezy.com/my-orders.
AI that works the moment you subscribe. Or use your own AI key, or the free on-device model.
There are three ways to run the AI. The free on-device model answers on your own PC and costs nothing to run; it is smaller and slower than a cloud model. With your own AI key you pay the AI company directly for what you use. The $12/month AI plan is for people who want AI to just work: no key and nothing to set up, one bill, 5,000 fast AI actions a month, 1,200 minutes of cloud dictation, and every Pro feature included while it is active. On-device dictation and transcription are unlimited with Pro or the AI plan.
Before you ask
Questions, answered.
Does the OCR need the internet?
No. Text extraction runs on your machine, with nothing to configure and no per-scan cost. The AI cleanup pass that tidies copied text is the only step that calls out. It sends only the text it read, never the picture, and you can turn it off in Settings, Screen capture.
Can it read a screenshot somebody sent me?
Yes. The overlay captures whatever is on screen, so a screenshot pasted into Slack, a PDF page that will not let you select, a video frame, or a scanned document all work the same way: drag over it, press Copy text.
Does it work on multiple monitors?
Yes, including monitors running at different scaling. KeptInFlow captures the desktop in the same coordinate space Windows reports the cursor and window positions in, so a selection on a 150% screen next to a 100% screen lands exactly where you drew it.
Where do my screenshots go?
Into a local capture library on your disk, indexed by their text so you can search them. There is no account and no upload. You can copy, save, drag out, or share any capture from there, and searching is done entirely on this PC.
Can I blur sensitive things before sharing?
Yes, by hand with the redact tool, or with a one-click pass that finds emails, card numbers, and keys in the capture on this PC and blurs them for you. The redactions are burned into the image before it is read or shared, so a blurred key stays blurred.
Is this a subscription?
No. KeptInFlow is a one-time license: the app is yours forever. The first 100 buyers get every update for life; later buyers get a year of updates, and renewing is optional ($39/yr). If you don't renew, you keep the last version you received. A capture-tool subscription stops working the year you stop paying.
Do I need an API key for the capture features?
Not for capture, OCR, annotation, recording, or searching your library. Those run on your machine. A key (or the 100 free AI actions a month) is only needed for the AI parts: the Ask chip, OCR cleanup, and the writing tools.
What about macOS?
Windows 10 and 11 only for now. That is where the deepest capture integration is possible today. macOS is on the roadmap but not dated; the license will cover both when it arrives.
Snip it. Read it. Find it later.
If Windows SmartScreen asks about this new release, click More info → Run anyway. The installer is code-signed and verified.