Privacy

What the website and browser extension read, send and keep.

Updated 20 November 2026. What came of it? is operated by James Pacileo.

Reading the site

No account is needed. We do not add analytics, set tracking cookies, keep reader profiles, or sell information about readers. Searching the published pages happens in your browser. A marker in your browser's session storage stops the wordmark animation repeating in the same tab; it is not sent to us.

Cloudflare hosts and protects the site. Requests for pages and files pass through its network, which receives information such as your IP address and browser request details and counts requests and errors. Our application's request logging is disabled. Cloudflare's processing is described in its privacy policy.

Opening the extension

When you open the extension, it reads the current tab's address and compares it with a public lookup file inside your browser. A reviewed match is answered without sending that page address. An untracked public report may send a cleaned page address to check for a previous automatic result. It downloads the lookup file from staging.whatcameofit.com and saves the file and its retrieval time in the browser, refreshing it after six hours. If the site cannot be reached, the saved copy can still be used. Removing the extension removes this local cache.

The extension also keeps two lists for you, in the browser only: the ten pages you checked most recently (each page's address and title, the date, and what the check found), and the pages you chose to follow, as they stood when you last looked. Neither list is sent to us or to anyone else. "Clear this list" empties the first, "Stop following" removes a page from the second, and removing the extension removes both.

While the extension is installed, twice a day it fetches the same public lookup file, so it can tell you with a notification when a page you follow has moved. That request carries no page address, nothing you checked and nothing about you. A reminder you set ("remind me in six months") is kept in the browser with the page's name and the day, shows one notification on that day, and is then removed. The browser may ask you to allow notifications; without them, the popup still marks a page that has moved.

Before the first untracked lookup, the extension asks you to allow sharing page content and screenshot fallback with our service and OpenAI. If you allow it, opening an untracked page starts the registry search described below automatically. You can turn this off in Automatic search settings and keep using matching against the published pages. The extension does not monitor other tabs or automatically search pages while you browse with it closed. Open it only on public reports you want to check. Select words before opening it to limit the text sent.

If a page contains visible images, video or a canvas and its text is insufficient to identify the report, opening the extension automatically captures the visible content. It prefers the largest visible media area, otherwise the viewport. It sends one reduced image, the caption and the cleaned page address through our service to OpenAI. A video contributes only its current frame, not its audio. Screenshots may include personal information visible on the page: open the extension only on public reports. Private browser windows and visible password or editing fields, and embedded documents overlapping the capture area, disable screenshot capture. Selecting words prevents automatic screenshots. Nothing is captured while the extension is closed.

The extension keeps up to ten successful automatic results in this browser for 24 hours, alongside a fingerprint of the content, so reopening a report need not make another paid lookup. A visual cache check may take a fresh local screenshot to compare its pixels, without sending it when the result matches. It does not save screenshot pixels or submitted page text. Previous results show their checked-at time. Search again requests a fresh lookup; Clear this list in Recently checked also clears these saved results. Removing the extension removes them.

Optional registry searches

On the website, a registry search runs when you press its search button. It sends the address or text you supply to our Cloudflare service. The service may read the article, a public post and its linked article, and sends the text to OpenAI to identify the treatment and condition being reported. It searches ClinicalTrials.gov, Europe PMC and OpenAlex using treatment names, the reported condition and research identifiers, then uses OpenAI to compare the trial records. The result is marked as an automatic, unreviewed search.

Organisation names explicitly mentioned in the report are sent toROR to find their public registry records. OpenAlex receives only treatment and condition search terms, and ROR receives only organisation names; neither receives your screenshot, page address or full article text. Public paper and organisation metadata, including empty search results, may be cached for seven days under a hash of the search terms. No abstracts are stored or shown. See OpenAlex's privacy policy.

In the extension, opening the popup on an untracked page reads the page or your selected words and sends that text and outgoing links to our service. A selection sends only your selected words, without other page links or a page address. A post also identifies its platform. Up to three X short links may be opened through t.co without cookies or a referrer. When no words are selected, a cleaned page address is included to check for an earlier result. Tracking parameters and fragments are removed; numeric article identifiers are preserved. Addresses with other query parameters are omitted. The text you choose may itself contain personal information. Please use public reports and avoid sending private medical or other sensitive information.

Our application does not log submitted content or store submitted text or screenshots. For a report the service can independently read publicly, it may retain the structured automatic findings for 24 hours, indexed by a hash of the cleaned address, and reuse them for other readers. These findings remain unreviewed and show when they were checked. Browser-only content, selected text and screenshots are not published to this shared cache. Your IP address is used by Cloudflare to limit the number of searches. We request that OpenAI does not save chat completions for later retrieval. This does not mean zero retention: OpenAI may keep prompts and responses in abuse-monitoring logs for up to 30 days, or longer for legal or safety reasons, and may temporarily cache model inputs. API data is not used to train models by default. Processing may take place outside your country. SeeOpenAI's data controls.

Automatic search has a limited monthly allowance. When it is used, the published pages remain available. A search never creates a page or changes a status.

Writing to us

Suggestion and correction links open your mail application. Nothing is sent until you send the email. Cloudflare forwards it to our Google Workspace inbox. Your address and message are used to handle your request, not to subscribe you to a mailing list. Please avoid including private medical records or other information that is unnecessary for the request.

For privacy questions, or to ask about access, correction or deletion of a message you sent, write to corrections@whatcameofit.com.