Browser Extension Privacy Policy
Last updated: September 4, 2026
This policy covers the Image to Prompt browser extension for Chrome and Edge. The website is covered separately by our Privacy Policy.
We have written this in plain terms rather than legal boilerplate, because the honest version is short.
What the permission at install is for
When you install the extension, Chrome says it can read and change your data on all websites. That permission exists for one feature: a small "Prompt" button that appears over the image under your pointer, on any page, so a prompt is one click away. It is described under "The button on images" below, and it can be switched off in the panel.
The extension does not read your browsing history, your cookies, page text, or anything you type. The script it runs on pages looks only at the image element under your pointer, and nothing leaves your browser until you act.
There is no advertising identifier and no third-party tracker in the extension. The only usage data it reports is described under "Usage counts" below.
What is sent, and only when you ask
Nothing leaves your browser until you act — a right-click on an image, a file you pick, an image you paste, or a link you enter. When you do, we send:
- The image. Either its URL (for a right-click or a pasted link, which our server then fetches) or the image itself, resized in your browser first (for a file you pick, drop, or paste).
- Your settings. The target model, output language, and depth.
That goes to our servers, which pass the image to our AI provider to produce the prompt.
The button on images
A small script runs on each page you visit and does one thing: it watches where your pointer is and, when it is over an image, draws a "Prompt" button in the image's corner. It reads nothing else from the page, keeps nothing, and sends nothing anywhere until you click the button — at which point the image is sent exactly as it would be from a right-click, including the fallback described above: if our server cannot fetch that image, the extension photographs the part of your screen showing it and sends that instead.
The switch "Show a button on images, on every site" in the panel turns the script off; it then stays idle on every page. With it off, clicking the toolbar icon or pressing its shortcut shows the buttons on the current page only.
Your clipboard
By default the extension copies the prompt to your clipboard as soon as it is ready, so it can be pasted straight into your image tool. It only ever writes to the clipboard; it reads it only when you paste into the panel yourself. The "Copy the prompt as soon as it is ready" switch in the panel turns this off.
Usage counts
So we know how many people use the extension each day and which parts of it get used, it sends small anonymous pings to our own analytics server (a self-hosted Umami instance) on five occasions: when you open the side panel, when a prompt is generated, when you bring a prompt back from the Recent list, when you switch the button on images on or off, and when you rate a prompt with the thumbs buttons.
Every ping carries the extension version and your browser's UI language. A generation ping also carries the model, output language and depth you chose, whether the image was handed over as a link or as a file, which way in you used (the right-click menu, the panel, or the button on an image), and the outcome as a short code — "ok", or an error code such as "QUOTA_EXCEEDED". A rating ping carries your vote and the same settings; a recall ping the model and depth; a switch ping whether it is now on or off. No ping ever includes the image, its URL, the prompt text, the page you were on, or any account or device identifier.
A thumbs rating is also sent to our own server, where it is stored against the prompt it is about, so we can tell which models and settings produce prompts people are happy with.
Umami sets no cookie and does not store your IP address; it uses the address and browser version only to tell one day's visitors apart, and keeps the counts.
What we store
- The generated prompt and its settings, so features like history work.
- An anonymous identity. On first use the extension creates an account with no email and no personal details attached, purely so a daily allowance can be counted. If you later sign in, that history moves to your account and the anonymous record is deleted.
- Your session token and your last used settings, in your browser's own extension storage. Signing out removes the token.
- Your last twenty prompts, each with a small thumbnail of its image, in your browser's own extension storage — the "Recent" list in the panel. They never leave your browser from there; the panel can remove any of them, or all of them.
We do not store the source image. It is passed through for the request and not retained.
Signing in
Signing in is optional and only raises the daily allowance. It happens on our website, not inside the extension — the extension never sees your password, your Google credentials, or any other secret. It receives a single-use token which it trades for a session, the same kind of session the website uses.
Who else sees the data
- Our AI provider, which receives the image in order to describe it.
- Our hosting and database providers, which store what is listed above.
- Our own analytics server, which receives the usage counts above. It is run by us, not by a third party.
We do not sell your data, and we do not share it for advertising.
Your choices
- Uninstalling the extension removes everything it stored in your browser.
- Signing out removes the stored session immediately.
- To have an account and its generated prompts deleted, write to support@mail.imagetoprompt.org.
Changes
If this policy changes in a way that affects what we collect, we will update the date at the top and, for a material change, note it in the extension's release notes.