About Markup
Markup is a free tool for drawing on a screenshot before you hand it to an AI model. Paste, point at the thing you mean, copy it back out.
What it does
You paste a screenshot into the page. You mark it up with arrows, boxes, numbered pins, freehand strokes and text labels. You copy the annotated image back to your clipboard and paste it into whatever you were doing.
That is the whole tool. There is no account, no upload, and no editing session to manage. The screenshot is decoded in the page, drawn on a canvas, and copied back out as a PNG. It never leaves your browser.
Why marked-up screenshots get better answers
A model reads an entire screenshot at once. It has no idea which part of it you are talking about until you tell it, and the words available for pointing are usually bad ones: the button, the second row, the thing on the left. On a dense interface, those phrases can match a dozen elements.
A red arrow does not have that problem. Neither does a numbered pin, which gives you a name you can use in the prompt itself: “Pin 2 wraps to a second line at 320px—why?” You spend one sentence on the question instead of three on describing where to look.
This matters most when you are asking about layout, spacing, alignment or state, where the difference between two elements is visual rather than textual, and where a description in words is exactly the thing that loses the information.
The tools
| Tool | Key | What it is for |
|---|---|---|
| Arrow | A | Pointing at one element without covering it. |
| Box | B | Scoping a region: “everything in here”. |
| Free draw | P | Circling something awkwardly shaped. |
| Text | T | A short label on the image itself. |
| Number | N | Numbered pins that advance as you click, so you can refer to them in the prompt. |
| Eraser | E | Click or drag over a mark to delete it. |
| Undo | ⌘Z | Redo with ⇧⌘Z. |
| Copy | ⌘C | Or just Enter. Puts the flattened PNG on your clipboard. |
Five colours, red by default because it survives almost any background. Light and dark themes follow your system and can be overridden.
How to use it
- Take a screenshot and copy it (⌘+⇧+Ctrl+4 on macOS, Win+Shift+S on Windows).
- Open markup.simplicated.dev and press ⌘V.
- Mark the parts you are asking about, press Enter, and paste into your model.
There is also a macOS app — the same tool in a standalone window, signed and notarized, working with no internet connection at all.
The guide goes further: which mark to use for which kind of question, and the prompt patterns that pair with them.
Nothing is uploaded. There is no server to upload to. The page makes no network requests once it has loaded. See privacy for the detail.
Frequently asked questions
What is Markup?
A free browser tool for annotating screenshots with arrows, boxes, numbered pins and text, aimed at people who are about to paste that screenshot into an AI model.
Why annotate a screenshot before sending it to an AI model?
Because the model sees the whole image and cannot tell which part you mean. Marking the exact element removes the guesswork, and numbered pins give you labels to refer to in the prompt instead of describing locations in words.
Do my screenshots get uploaded anywhere?
No. Everything runs in your browser. The image is decoded, drawn on a canvas and copied back to your clipboard locally. There is no server, no account and no analytics.
Which AI tools does it work with?
Any of them that accept an image: ChatGPT, Claude, Gemini, Copilot, Perplexity, or an API call you write yourself. Markup produces an ordinary PNG, so nothing about it is specific to one model.
Is it free? Do I need an account?
Free, and no. There is no sign-in, no trial and no paid tier.
What can I paste into it?
Any image your browser can decode, including PNG, JPEG, WebP and GIF. You can also drop a file onto the page or click to browse for one.
How do the numbered pins work?
Each click drops the next number, so a click sequence gives you 1, 2, 3 without changing tools. If you erase or undo a pin, the counter follows, and the next click reuses that number.
Can I move or edit a mark after drawing it?
Not yet. Marks are committed as you draw them; to change one, erase it and redraw. Undo (⌘Z) steps back through them in order.
Does it work on a phone or tablet?
Drawing does, with touch or a stylus. Pasting from the clipboard is the weak spot on mobile browsers, so on a phone you will usually pick the image with the file browser instead.
Keep reading
- How to mark up a screenshot for an AI model — which mark to use for which question, plus prompt patterns.
- Privacy — what the page stores and what it does not.
- Open the tool — nothing to install.