The short version
You can read every page here without an account, and without being counted unless you choose to be. There is no account to make, no comment box, no newsletter, and no form that asks for a detail about you. The one address on the site is written out for you to copy rather than wired into a form, so nothing is submitted anywhere by clicking it.
Two things are kept, and one of them is your answer. A web server access log, the same thing almost every server on the internet writes, whose exact format section 02 prints rather than describes. And, only if you allow it, Google Analytics, which section 04 covers on its own because it is the one thing here that involves somebody else.
This page will get longer if that stops being true. The honest version of a privacy page on a site with no measurement is a short one, and the temptation is to pad it with clauses about things that do not happen here.
What the server writes down
One line per request, in this format, which is the nginx default with nothing added to it:
$remote_addr - $remote_user [$time_local] "$request" $status $body_bytes_sent "$http_referer" "$http_user_agent" "$http_x_forwarded_for"
| Field | What it is | Does it identify you |
|---|---|---|
| remote_addr | The address the request arrived from, which is this site’s own proxy rather than yours, because no real_ip directive is configured |
No |
| x_forwarded_for | The address the proxy says the request came from, which is yours | Yes, and it is the only field that does |
| time_local | When the request arrived, to the second | No |
| request | The method and the full path, so a search shows the words you searched for | Only if you typed something about yourself into the search |
| status, body_bytes_sent | What the server answered and how large the answer was | No |
| http_referer | The page you came from, when your browser chooses to send it | No |
| http_user_agent | The browser and operating system string your browser sends to every site | No |
How long it lasts
The log is not kept by date. It goes to the container’s output stream, which the host caps at three files of ten megabytes, so thirty megabytes at most and the oldest line falls off the end when a new one arrives past that. There is no archive, no copy taken before it rolls, and nothing exported.
Which means the retention period is a volume rather than a promise about time, and that is stated here because rounding it to a tidy number of days would be a guess.
What it is for
Reading it when something breaks. A page answering the wrong status code, a link that turns out to be dead, a crawler taking the site apart faster than it should: the log is the only way any of those become visible.
It is not read to count visitors, because it would be a bad way to do that and because nothing here wants the number badly enough to start.
What is not here, and how to check
Every row below is checkable from your own browser, which is the only kind of privacy claim worth making. Open the network panel and load any page.
| Not here | What that means | Check it |
|---|---|---|
| No cookies until you allow one | No page sends a Set-Cookie header. Allowing analytics is what creates the first two, and turning it off again expires them |
The storage panel, before and after |
| No analytics until you allow it | The tag is not fetched, no dataLayer exists and nothing is sent until there is a stored yes. A no is remembered and not asked again |
The network panel: no Google host |
| One domain, unless you allow the second | Every font, image and stylesheet comes from this domain, and with analytics declined a page loads nothing from anywhere else. Allowing it adds Google and nothing more, which the policy header restricts rather than trusts | The network panel: count the hosts |
| No advertising | No ad network, no affiliate link and no sponsored record, which the about page also states as a funding position rather than a privacy one | the method page |
| No fingerprinting | Nothing this site writes measures your fonts, your canvas or your timing. What Google Analytics collects is listed in section 04 rather than paraphrased here | The scripts the page loads |
| No social embeds | No like button, no share widget, no video player from anywhere else | Any record page |
The analytics, and your answer
The measurement is Google Analytics 4, and it runs only after you allow it. Until then the tag is not fetched, there is no dataLayer on the page and no request reaches any Google host: a browser that has declined, or that has not been asked yet, talks to this domain and to nothing else. That is checkable in your own network panel and it is the whole reason the bar exists.
The question it is there to answer is which record to write next. 48 records exist and the queue on every silo page is longer than what is written. Reading which pages people actually open is a better way to order that queue than guessing, and it is the only use this data has here.
| Item | What it is | Where it lives |
|---|---|---|
| A random identifier | A number generated in your browser so a second page view can be recognised as the same visit. It is not derived from anything about you and it means nothing outside this site | The _ga cookie, and a second one for the session |
| Which page, and where from | The address of the page and the referring page, which is the part that answers the question above | Sent to Google |
| Browser and device | Browser, operating system, screen size and language, the same strings your browser sends to every site it visits | Sent to Google |
| Approximate place | A country and usually a city, worked out from your network address at Google’s end | Sent to Google |
| Your network address | Used to derive the place above and, Google states, not stored or reported. We cannot see it in the reports and cannot verify what happens to it beyond what Google publishes | Not visible to this site |
| No advertising signals | Ad storage, ad personalisation and ad user data are all set to denied, because there is no advertising here and nothing to remarket | Refused in the configuration |
Change your answer, now
Whichever you chose, it is stored in this browser and nowhere else, so it is not something we hold about you. Turning it off also expires the two cookies rather than leaving them behind.
Your choice is stored in this browser. The buttons above change it either way.
And Cloudflare, which is in front of all of it
Every request to this site reaches Cloudflare before it reaches the server, because that is what serves the certificate and the address. So Cloudflare sees the request whatever you answer above, and that is not something a choice on this page can turn off: it is how the site is reachable at all.
Cloudflare also tries to inject a measurement script of its own at that edge, which our pages do not contain and the content security policy refuses. It has never been allowed to run here, and it is being switched off at the source rather than left to be blocked every time.
Search, and where the words go
It never leaves this server
Search here runs on the server rather than in your browser. What you type is sent to this site, which asks its own search index, on a private network with no route to the internet, and returns the page. No search service outside this machine sees the query, because there is no search service outside this machine.
The index software ships with usage reporting switched on by default. It is switched off here, in the configuration, so it sends nothing to its own authors either.
What that leaves
The query is part of the address of the results page, so it appears in the access log line described in section 02, beside your network address. That is the one place a search term is written down, it is subject to the same rolling limit as everything else in that log, and it is not read.
If that matters for a particular search, the records are all reachable without searching: the A to Z index lists every one of them.
If you write to us
The address on the about page is an ordinary mailbox. If you write to it we have whatever you put in the message, including your address, for the same reason anybody has an email you sent them.
What happens to it: it is read, it is answered, and a correction that changes a record gets that record changed. Your message is not added to a list, and there is no list. Nothing is forwarded to anybody, and no reply carries a tracking pixel, which is worth saying because a great many replies do.
A correction does not need your name to count. The thing that decides whether a figure changes is the source you cite, so if you would rather write without identifying yourself, that costs the correction nothing.
If any of this changes
What already changed it
Measurement, in September 2026. This page predicted two things that would change it and one of them arrived the same week it was written: Google Analytics, behind the bar, for the question in section 04. The other, equipment affiliate links, has not, and the method page still carries the rules it would have to follow.
What did not change is the funding. No advertising, no affiliate link and no sponsored record, which is a separate promise from this one and is still whole.
How you found out
Exactly the way this paragraph promised: the month is named above, and the four rows in section 03 that stopped being true were rewritten rather than quietly softened. A privacy page that keeps a claim after it stops holding is worse than not having the page, and the test of that is whether it happens on the day the claim breaks.
There is no version history here to point you at yet. Until there is, the rows above are the commitment, and every one of them is falsifiable in your own browser in about ten seconds.
Questions about privacy here
Do you use cookies?
Only the two Google Analytics sets, and only after you allow it. No page here sends a Set-Cookie header of its own, and your answer to the bar is kept in local storage rather than in a cookie, so declining leaves you with none at all. Turning analytics off later expires the two.
Do you have analytics?
Google Analytics 4, from September 2026, and only for visitors who allow it. Before that this page said there was none, which was true when it was written and is the reason section 07 explains what changed instead of pretending the page always read this way.
Do you log my IP address?
Yes, in one field of one log line, and it is the only field in it that identifies you. It is capped at thirty megabytes of rolling log with no archive behind it, it is not joined to anything, and section 02 prints the whole line so you can see exactly what sits beside it.
Are my searches recorded?
The query is in the address of the results page, so it is in that same log line. It never reaches a search service outside this server, because the index runs on the machine and its own usage reporting is disabled in the configuration.
Do you sell or share anything?
There is nothing to sell. No profile is built here and the log is not a product. Google Analytics assigns a random identifier if you allow it, which section 04 describes, and Google is the only third party involved in any of this. No third party has access to this server or to anything on it.
What about the sites you link to?
Every figure on a record names its source and links to it, and following one of those links takes you to somebody else’s site under their rules rather than these. That is the trade this site is built on: the citations are the point, and they lead off the domain.
Can I ask you to delete something about me?
Write to the address on the about page and it will be done. In practice there is very little to delete, because there is no account and no profile, and a log line is gone on its own once thirty megabytes of traffic have passed over it.
Is this a legal document?
It is a description of what the server does, written to be checked rather than to be filed. If you need something in a particular legal form for a particular jurisdiction, write and say which, and it can be stated properly instead of guessed at here.