Find every photo of yourself online with Claude (audit + takedown kit)
Hey, it's Cindy 🌱 By the end you will have a tracker of every place your face shows up online, plus ready-to-send takedown requests for the ones you want gone.
btw the waitlist for AI System Club is now open! If you're keen on using AI to automate your content, video editing, job search, business or build AI agents, feel free to sign up (only if you want to hehe). Ps. the waitlist gets first dibs:)
One honest note first. Claude will not run the face search for you, and that is by design. Anthropic's usage policy prohibits facial recognition, and Claude in Chrome is explicitly blocked from gathering or scraping facial images, so it will refuse. So you run the engines, which takes about five minutes, and Claude does everything around them: it organises every hit into a tracker and writes the takedown letters. That is still the slow part, and it is the part you would otherwise never finish.
start here
The 5-step audit
🙂 Pick one clear selfie. Face forward, good light, no sunglasses. This is the only input the engines need.
🔎 Run it through the engines yourself. PimEyes, FaceCheck ID and Yandex Images. This is the manual bit, and it takes about five minutes.
🔗 Copy every hit into Claude. Paste the links you found straight into a Claude chat, one per line, with the site name if you have it.
🗂️ Build the exposure tracker. Every hit lands in one table: link, site, and how exposed you are, ranked worst first.
✉️ Send the takedown letters. Claude writes a tailored removal request for each one, citing the law that applies to you.
do this first
Step 1: run the engines yourself
Upload the same selfie to each engine below and collect every result that is actually you: the page link, the site name, and what the photo or page is. Keep it in a scratch note, you are about to paste the whole lot into Claude in one go. Ignore the obvious false matches, and do not chase results that are clearly someone else.
copy these
The 3 prompts
First, hand Claude the hits you collected in step 1:
🔗 the handover prompt
I am running a privacy audit on myself. I have already run my photo through the face-search engines and these are every page where I appear:
[paste your links here, one per line]
For each one, open it and tell me: the site, what the page actually is, whether it looks like a real match or a false positive, and who the right contact is for a removal request (a privacy email, a report form, or a data-broker opt-out page). Ask me before you assume anything you cannot see on the page.
Then, once the hits are in, organise them:
🗂️ the exposure tracker
Put every hit in a tracker with the link, site, and how exposed I am.
Make it a table with columns: site, link, what it is, exposure level (low / medium / high), removal route (site contact, platform report form, search-engine removal, data-broker opt-out), and status. Rank it by exposure so I handle the worst first, and flag any that look like a false match I should double-check before I act.
Finally, get it removed:
✉️ the takedown letters
Write a removal request for each high-exposure entry, citing their rules and my right to erasure.
Tailor each one to that specific site: reference the exact URL, state that I am the person pictured and did not consent to publication, request removal, and cite the strongest basis that applies to me: GDPR Article 17 if the site serves the EU or UK, the California CCPA if I am in California, or the Australian Privacy Act if I am in Australia. Give each one a subject line and keep it under 150 words.
the engines
Where the matching actually happens
You run these yourself, so it helps to know what each one is for. General search engines are weak on faces by design, so the face-specialised ones do the real work. Run all three if you can, they surface different things.
PimEyes The most thorough. Free to see that a match exists, but the source links sit behind a paid plan, so start here only if you are willing to pay to see where.
FaceCheck ID Face-specialised across social profiles, blogs, news and video. Free preview of results, and it scores each match, so treat low scores as maybes.
Yandex Images The best genuinely free face-matcher, and it often surfaces hits the others miss. Weaker on Western social platforms since 2022.
Skip Google Lens for this It will find pages containing your exact image, but it will not identify a private person by face, and that is deliberate. Useful as a fourth pass, useless as your first.
The takedown letters lean on real privacy law, and there is one for wherever you are, so Claude cites the strongest fit.
GDPR Article 17 The EU and UK "right to erasure": you can ask a site to delete your personal data. Full text: Article 17.
California CCPA / CPRA Gives California residents the right to request deletion of personal information a business holds.
Australian Privacy Act The Australian Privacy Principles let you ask an organisation to correct or stop holding your personal information.
The site's own rules Most platforms have a non-consensual imagery or privacy policy. Citing it alongside the law makes a removal request much harder to ignore.
🔁 Removing a photo from a site does not instantly clear it from search results, so recheck the engines a couple of weeks after a takedown, and re-run the whole audit every few months. Your tracker turns the re-run into a five-minute job.
the waitlist is now open
AI System Club is launching soon
Build AI systems with me through recorded lessons you can follow at your own pace. Join the waitlist for first access and the founding price.
btw this list gets first dibs the moment doors open :)