Privacy Policy
Last updated: 26 July 2026
Clunky AI is built around public website scans, emailed reports, editorial publishing, and benchmark reporting. We collect the minimum personal data needed to run those surfaces.
What we collect
- URLs submitted for scans and the raw signals produced by a scan.
- Email addresses submitted for reports or scan follow-ups.
- Names, email addresses, company names, website URLs and messages submitted with a Clunk Removal or AI Visibility Sprint inquiry.
- Optional industry and builder information submitted with a scan.
- The normalised domain submitted to the llms.txt generator or checker, together with the tool action and outcome.
- Hashed IP addresses for abuse prevention and dataset integrity.
We do not store raw IP addresses. IPs are hashed with a server-side salt before storage.
How we use scan data
Every scan contributes to the Clunky Index. We retain scan rows indefinitely so the benchmark can track methodology versions and trends over time. Public rollups are aggregated; we do not publish personal email addresses.
Report delivery and scan follow-ups are handled separately from the scan itself. You can unsubscribe from future follow-up emails with one click.
Service inquiry details are used to assess the requested work, reply to the inquiry and keep a record of the proposed scope. They are not added to marketing lists unless you separately opt in.
Analytics
The default analytics posture is privacy-first and cookieless. We use Vercel Web Analytics for anonymous, aggregated page-view measurement. Dynamic report identifiers and query strings are removed before page views are sent. Product-event plumbing can also use Plausible when it is configured. For the llms.txt tool, the normalised domain is attached to generation, checking, copy and download events so we can understand which websites use the tool. We do not send the full submitted URL, fetched page content or generated file to analytics. Website URLs entered into the general scanner or service inquiry are not sent to either analytics service.
After a successful server-side llms.txt generation or published-file check, we send Clunky an internal operational email containing the normalised domain, action and summary result counts. It does not contain the submitted path, fetched page content or generated file. Pasted files are checked entirely in the browser and do not trigger a notification.
Processors
Depending on environment configuration, we may use Supabase, Resend, Vercel, Fly.io or Railway, Google PageSpeed Insights, OpenAI for narrative text, and Plausible. LLMs are not used to produce numeric scores.
Your rights
If you are in the UK or EEA, you can request access, correction or deletion of personal data such as your email address. Scan rows may be retained in anonymised or aggregated form to protect benchmark integrity.
Contact
Email: privacy@clunky.ai