Privacy Policy

CreepJS is an educational, privacy-first fingerprinting platform. This page explains what data is processed, what is stored, and how you can opt out.

Site: https://creepjs.org • Contact: hello@creepjs.org • Terms

What CreepJS Does

CreepJS demonstrates how browser fingerprinting works by running collectors in your browser (Canvas/WebGL/Audio/Navigator/Storage APIs and more) and presenting the results in a human-readable format.

The project is intended for education, security research, and legitimate use cases (e.g. fraud prevention). You should always follow applicable laws and disclose fingerprinting usage in your own products.

The Terms of Use define the project's responsible-use boundaries.

Data Collected & Stored

In the browser: fingerprint components are computed locally to generate a fingerprint ID and to render the on-page report. The homepage and the checker start a scan when they load; elsewhere a full scan runs when you start it. Its WebRTC check contacts Google's public STUN servers (stun.l.google.com), which see your IP address the same way any website does.

Checker history: the checker keeps your last 50 scans in this browser's localStorage so you can compare them. They never leave your device, WebRTC addresses and media device labels are removed before saving, and “Clear All” deletes them.

On the API: the API is designed to be stateless with respect to fingerprint payloads. It validates requests and returns responses, and it maintains minimal operational state for abuse prevention.

  • API tokens are stored in Cloudflare KV with the email address they were requested for and their creation time.
  • Daily usage counters (per token owner, per visitor, and for token requests) are kept in a Cloudflare Durable Object under SHA-256 hashes of the email address or visitor IP address, never the raw value, and reset after 24 hours.
  • IP lookups (the public IP checker, the token API, and the optional “My IP” risk check) cache the provider's response in KV under the looked-up IP address for up to 24 hours (1 hour for high-risk results).

Optional population comparison: the checker sends a Canvas hash, WebGL renderer, timezone, font count, and screen tuple only when you activate “Compare with population.” Small matching groups remain suppressed. This is separate from submitting a full report in the API playground.

CreepJS does not intentionally store raw fingerprint component payloads long-term.

Analytics

When enabled, Cloudflare Web Analytics provides aggregate page-view metrics without analytics cookies. Cloudflare hosting security can separately set a cf_clearance cookie for automated-traffic checks.

Product events are sent to our own API. Event storage is currently disabled for this deployment, so valid event payloads are discarded. When event storage is enabled, Cloudflare Workers Analytics Engine retains aggregate events for up to three months. Events describe tool runs (checker scans, IP risk lookups, playground API requests), API token requests, docs section views, fingerprint-ID copies, and clicks on a few tracked buttons and links, including partner-card impressions and partner-link activations.

The event format permits only:

  • the event name;
  • the page path, without query string or fragment (so an IP you looked up via ?q= is never stored);
  • a few fixed labels: which tool and page section, success or error, and the HTTP status code of the tool's API call. Partner events use the provider ID and placement, such as the IP result or directory.
  • No IP address, user agent, cookie, email address, API token, fingerprint ID or fingerprint data is stored with an event.
  • Events cannot be linked to each other or to a visitor: there is no visitor or session identifier.
  • Product events are not sent when Do Not Track or Global Privacy Control is enabled in your browser.

If your jurisdiction requires consent for analytics or fingerprinting, you must add that consent layer in your own deployment before enabling those features.

Partner recommendations

We may earn a commission from qualifying purchases through disclosed partner links. Listings use reviewed public information and a fixed editorial order. IP addresses, risk scores, and fingerprints do not select providers, change results, or enter partner links.

Partner impressions and link activations are deduplicated in memory for each page instance. This does not create cookies, a visitor ID, or persistent advertising storage. DNT and GPC suppress these events. Observed clicks are not orders or unique visitors.

Partner links use a no-referrer policy and only reviewed, fixed attribution parameters. We do not prefetch merchant pages or load merchant pixels. Following a link takes you to a third-party service, which receives an ordinary visit and applies its own privacy policy. Orders, refunds, and settlement are confirmed in merchant reports; we do not match individual visitors across sites.

Your Choices

The homepage and the checker start a full scan on load; the fingerprinting guide waits for you to start one. Individual signal checks run automatically unless they need the external WebRTC network check. You can read documentation and policies without starting a full scan.

To remove stored checker scans, open local scan history and choose “Clear All.” The playground keeps tokens only in this tab’s memory; reload it to clear a token. Copying or sharing an IP report link shares the queried address.

Signals like Do Not Track and Global Privacy Control are part of the privacy posture, but they do not replace your own product-level compliance review.

For background and technical details, see the documentation and the browser fingerprinting guide.