Documentation

Everything you need to know about using ScreenToPrompt to annotate screenshots and generate AI-ready prompts.

Getting Started

ScreenToPrompt is a Chrome extension that lets you capture any webpage, annotate it with visual tools, and generate a precise AI prompt describing your changes. Here's how to get started:

Install the extension Add ScreenToPrompt from the Chrome Web Store. It's free.
Click the extension icon A popup appears with three capture mode options.
Choose a capture mode Current View, Full Page, or Selected Area.
Annotate your screenshot Use arrows, rectangles, text, comments, sketches, blur, markers, and more.
Generate an AI prompt Click Prompt to have AI analyze your annotations and generate a detailed instruction.
Paste into your AI tool Copy the text or image and paste it into ChatGPT, Claude, Cursor, Bolt, or any AI coder.
ScreenToPrompt extension popup showing capture modes

Capture Modes

When you click the ScreenToPrompt icon, you'll see three options:

The Toolbar

After capturing a screenshot, the annotation toolbar appears at the bottom of the screen. It contains all the tools you need to mark up your screenshot.

ScreenToPrompt annotation toolbar

From left to right, the toolbar contains:

  1. Comment — Add a comment with a text label attached to a rectangle area
  2. Arrow — Draw arrows to point from one element to another
  3. Rectangle — Highlight areas with rectangles (stroke, filled, or dashed)
  4. Text — Place text labels directly on the screenshot
  5. Sketch — Freehand drawing for circles, lines, and X marks
  6. Blur — Pixelate sensitive areas like passwords and API keys
  7. Erase / Cover — Cover areas with a matching background color
  8. Number Marker — Place numbered crosshair markers for step-by-step instructions
  9. Color Picker — Select the active color for your annotations
  10. Annotation Select — Click an annotation to select, move, resize, or edit it
  11. Screenshot Select — Select a region of the screenshot to move, copy, or delete. Click the dropdown arrow to switch between rectangle and freeform lasso modes.
  12. Undo — Undo the last action
  13. Clear All — Remove all annotations
  14. Markups / Layers — Open the layers panel to manage individual annotations
  15. Copy — Copy the annotated screenshot to your clipboard as an image
  16. Prompt — Generate an AI prompt from your annotations
  17. Close — Close the annotation editor

Some toolbar buttons have dropdown menus. For example, the Screenshot Select tool lets you choose between Rectangle and Freeform selection modes:

Screenshot Select dropdown with Rectangle and Freeform modes

Annotation Tools

When you select an annotation, a toolbox appears in the top-right corner with options specific to that annotation type.

Annotation toolbox with color, size, and style options

The toolbox lets you change color, adjust size with a slider, switch between stroke/fill/dashed styles (for rectangles), and delete the annotation. Click the back arrow to deselect.

Color palette with 8 swatches and eyedropper

Arrows

Arrows

Click and drag to draw an arrow from point A to point B. Great for showing where elements should move to. The AI interprets arrows as "move this element here" instructions.

Toolbar with Arrow tool selected

Rectangles

Rectangles

Click and drag to draw a rectangle around any area you want to highlight. Switch between stroke (outline), filled (semi-transparent), and dashed styles using the toolbox buttons.

Toolbar with Rectangle tool selected

Text

Text

Click anywhere on the screenshot to place a text label. Type your instruction (e.g., "remove", "bigger", "change to blue") and the AI will interpret it in context. Text supports up to 400 characters.

Toolbar with Text tool selected

Comments

Comments

Click and drag to draw a rectangle, then type your comment. The text appears directly below the rectangle. Use comments for detailed change instructions like "change this button to blue with rounded corners".

Toolbar with Comment tool selected

Curved Comments

Curved Comments

An advanced comment style with a curved bezier connector between the rectangle area and the text label. The text can be dragged independently, and the comment parts can be sub-selected like in Figma. Choose this for more precise placement control.

Sketch

Sketch / Freehand Draw

Draw freely on the screenshot. Circle elements to highlight them, draw an X to mark something for deletion, or sketch lines to show connections. The AI interprets common drawing patterns automatically.

Toolbar with Sketch tool selected

Blur

Blur / Pixelate

Click and drag to blur sensitive areas. Pixelates API keys, passwords, emails, and other personal data before sharing screenshots with AI. Essential for privacy when annotating settings pages.

Toolbar with Blur tool selected

Erase / Cover

Erase / Cover

Cover areas with a solid color to completely hide content. Unlike blur, erase fills the area with a flat color. Use the eyedropper to pick a matching background color for seamless coverage.

Toolbar with Erase tool selected

Number Markers

1
Number Markers

Click to place numbered crosshair markers. Each marker automatically increments the number. Use them to create step-by-step instructions where each number points to a specific element. The precision crosshair helps the AI identify exactly which element you're targeting.

Toolbar with Marker tool selected

Eyedropper

Eyedropper

Pick any color from the screenshot to match your design system. Available in both the main toolbar color palette and the per-annotation toolbox. Click the eyedropper button, then click anywhere on the screenshot to sample a color.

AI Prompt Generation

The Prompt button is the core feature of ScreenToPrompt. When you click it, AI analyzes your annotated screenshot and generates a detailed, structured prompt describing every change you've marked.

The AI understands visual context:

After generating, you have two copy options:

The prompt panel has two modes:

AI prompt panel in read mode with formatted text and edit button

In read mode, the generated prompt is displayed with formatted text. Click the pencil icon to switch to edit mode.

AI prompt panel in edit mode with textarea and save/cancel buttons

In edit mode, you can modify the prompt text freely. Click the checkmark to save or the X to cancel your edits.

AI Setup

To generate AI prompts, you need to connect an API key from one of the supported providers.

AI settings panel with provider selection and API key input
Choose a provider Select OpenAI, Anthropic, or Google (Gemini) from the dropdown.
Paste your API key Copy your API key from the provider's dashboard and paste it into the field.
Test & Save Click "Test Connection" to verify your key works, then click "Save".

To access AI settings, click the extension icon and expand the "AI Settings" section at the bottom of the popup.

Privacy: Your API key is stored locally in your browser using Chrome's built-in storage. It is never sent to ScreenToPrompt servers (we don't have any). When you generate a prompt, your browser communicates directly with the AI provider's API.

Keyboard Shortcuts

Speed up your workflow with these keyboard shortcuts while the annotation editor is open:

Action Shortcut
Undo Ctrl + Z / + Z
Delete selected annotation Delete / Backspace
Deselect / Cancel Escape
Nudge selected annotation
Large nudge Shift + Arrow keys
Copy annotated screenshot Ctrl + C / + C
Zoom in (full page mode) Ctrl + + / + +
Zoom out (full page mode) Ctrl + - / + -
Reset zoom Ctrl + 0 / + 0

Layers Panel

The Layers panel (click the Markups button in the toolbar) shows every annotation on your screenshot in a list, similar to the layers panel in Figma.

Layers panel showing annotation list

In the layers panel you can:

Each layer shows a thumbnail preview, the annotation type and content, and its dimensions.

FAQ / Troubleshooting

Why does the Prompt button say "Add API key in extension settings"?

You need to configure an AI provider API key first. Click the extension icon, expand "AI Settings", choose a provider, paste your API key, and click Save. See AI Setup for details.

Which AI providers are supported?

ScreenToPrompt supports OpenAI (GPT-4o), Anthropic (Claude), and Google (Gemini). All three can analyze screenshots and generate prompts. Pick the one you already have an API key for.

Is my data sent to ScreenToPrompt servers?

No. ScreenToPrompt has no servers. Your screenshots never leave your browser until you copy them. When you generate a prompt, your browser communicates directly with your chosen AI provider. Your API key is stored locally in Chrome's sandboxed storage.

How do I annotate a full-page screenshot?

Choose "Full Page" in the capture popup. The entire page will be scroll-captured and opened in a pan & zoom viewer. Use mouse wheel or the zoom controls in the toolbar to navigate. All annotation tools work the same way.

Can I edit an annotation after placing it?

Yes. Switch to the Annotation Select tool (cursor icon) or click the annotation directly. The toolbox will appear letting you change color, size, style, and delete. Double-click text or comments to edit the content.

How do I match a color from the screenshot?

Open the color palette by clicking the color circle in the toolbar. Click the eyedropper button at the bottom of the palette, then click anywhere on the screenshot to sample that color. The eyedropper is also available in the per-annotation toolbox.