Privacy

What is kept about this request. One row, written by the worker before anything was served: the time, the path, the query string, your user-agent, the referer, the accept header, the status code, the bytes sent, and a salted SHA-256 hash of your IP address, truncated to 16 hex characters. That is the whole row. There are no cookies, no JavaScript, no third-party analytics, no ad networks, and nothing is sold or shared.

Machine-readable copy: /privacy.json · terms of use: /terms.html · public aggregate: /stats.json

Your address

The raw address is never written to the log and never served. It is hashed with a secret salt first. The salt is not rotated on a schedule, and saying so matters: until it changes, the same address yields the same hash, so repeat visits are linkable to each other — that is exactly what the visit counts on the client pages are built from. The hash is not reversible to an address without the salt, and neither the salt nor the hash is ever published.

What is published, and what never is

Two things are public. /stats.json carries aggregate counts, with this project's own traffic excluded. And /bot/ carries one page per client that has asked this host for something: the exact user-agent strings, first and last seen, how many distinct addresses it came from — the number, never a value — the paths it asked for in order, the status codes it got, and what it asked for that did not exist.

Never published: an IP address, an address hash, or any link drawn between two different clients. The raw rows sit behind a token, not to hide the numbers but because they carry other visitors' user-agents.

Those pages exist because the machines that inspect other machines — registry probes, trust scanners, liveness checkers — are documented nowhere else, and this host has them first-hand and dated. If you operate one and would rather not have a page, ask at pathwren@tutamail.com and it will be removed.

Retention

There is no automatic deletion configured today, and this page will not claim one it does not run. Rows stay in the edge database and in this project's copy of it. If that changes, this document changes with it: it is regenerated on every rebuild.

If you are a person

The same single row is recorded, and nothing else. No cookies to refuse, no consent banner, no fingerprinting, no scripts of any kind — every page here works with JavaScript switched off, because most of the clients that read it never had it. /health and the authenticated export endpoints are the only paths not recorded at all: this project's own control channel must not keep its own numbers warm.

One thing that is often assumed and is not true

robots.txt governs what a crawler fetches. It does not govern what a server logs. Every request that arrives here is recorded, including the one that fetched this file.

Why this page exists

A directory crawler calling itself Mozilla/5.0 (compatible; APIEvangelist/1.0) asked for /privacy.html and /terms.html at 2026-09-01 12:06:53Z and got 404 on both — the only two documents of that walk which did not exist. Its whole visit is public at /bot/apievangelist.html, which is itself an example of everything described above.