Privacy

The service itself collects and stores no personal data: no accounts, no IP logging, aggregate counts only. What the service does NOT do:

• No accounts. No login and no registration; no email is required or collected to see results or to use the API under the free limit.
• No IP logging. The caller IP is not written to an access log (the service runs with the web-server access log turned off), not written by the rate limiter (its token buckets are held in memory only and never persisted), and not written by instrumentation.
• Only aggregate event counters. Instrumentation keeps one row per (channel, surface, event, day) with a count. It holds no IP, no hashed IP, no organisation key, no subject identifier, and no request metadata. The MCP surface is the same: one aggregate counter per (tool, event, day), no caller identifier of any kind.
• No cookies used for tracking.

Two narrow exceptions — stated honestly

The service surfaces process no personal data. Two narrow paths, both started by you contacting us, are the only non-zero cases, and neither keeps a personal record:

• Correspondence. If you contact us (for example to dispute an observation), we process your message solely to respond and delete it once the matter is resolved — never stored in a database or used for anything else.
• Dispute log. We keep a de-identified case record of each dispute (date, entity, what was disputed, our assessment, action, response date) for methodology traceability, and publish it at /dispute-log. It contains no personal data — no name, no email, no address, and not the message text.

What the subjects are

The subjects observed are software components — GitHub organisations, repositories, and packages — which are not personal data. The stored record is dated component state (HTTP status, last-commit date, archived flag) and the public source URL. Commit-author identities, package-author emails, and anything derived from version history are not stored.

Contact

Privacy questions: [email protected]. Disputes about an observation: [email protected]. These are role addresses; no personal mailbox is used.

The dormant user database

A user-database and consent model exists in the code but is inactive. No request route reaches it. Activating it is the point at which personal-data processing would begin; until then, none is.