Privacy
This site does not know who you are, and it is built so that it cannot find out. There is nothing to sign up for and nothing to fill in. Since 20 August it does set Google Analytics cookies, which is a change from how it worked before, and this page says exactly what they are.
Last updated 20 August 2026 · written from the code, not from a template
- No accounts, no sign-in, no newsletter, no comments — the site has no forms of any kind for readers.
- Two of our three measurement tools are cookieless. The third, Google Analytics, sets cookies named _ga and _ga_G-44CBEZQND1 to tell a returning visit from a new one.
- We never learn who you are: no name, no email, no account, and nothing you type is recorded.
- No advertising network, no remarketing, no session recording, no embedded video, no social widgets. Google Analytics is configured to count visits, not to build audiences.
- Fonts and images are served from this site. Reading a page does contact Google, because Google Analytics is on it.
- We never sell data, and we have nothing anyone would want to buy.
Who is responsible for this
AI Intelligence is an independent AI testing site published by Robert Prime, who is the data controller for everything described on this page. There is one mailbox, it is read by a person, and privacy questions go to the same place as everything else: corrections@aiintelligence.com — put Privacy in the subject line and it will be picked up. Who runs this site · Contact
What the site actually is, technically
Every page you can read here is static HTML, generated when the site is built and served from Vercel’s network. It is not an application you log into; there is no reader account, no profile, no saved preferences and no basket. The rankings, prices and receipts are files, and you are reading files.
There is one private area, at /admin, used by the site’s own editor to queue tests and record corrections. It is the only part of the site that authenticates anybody, and it is not reachable by readers.
What we measure, and with what
Vercel Web Analytics counts page views. We do it because a site that publishes tests needs to know whether anybody opens the evidence — if nobody ever clicks a receipt, the receipts are decoration and we should know that. This one is cookieless: Vercel’s own documentation states it works “without using any third-party cookies, instead end users are identified by a hash created from the incoming request”, and that “the lifespan of a visitor session is not stored permanently, it is automatically discarded after 24 hours”.
Everything it may record, in full. This is Vercel’s own published list, with their examples — there is no eleventh item we have left out:
- Event timestamp2026-08-17 09:06:30
- Page URL/best/writing
- Dynamic path/best/[category]
- Referrerhttps://www.google.com/
- Filtered query parameters?ref=hackernews
- Approximate location from the requestGB, England, London
- Device OS and versionAndroid 10
- Browser and versionChrome 86 (Blink)
- Device typeMobile, desktop or tablet
- Analytics script version1.0.0
Note what is not in that list: no name, no email, no account, no IP address stored against the record, and no identifier that could follow you to another website. Vercel documents the recording as anonymous and states it does not collect or store information that would let a browsing session be reconstructed across sites.
Five clicks are counted as named events, each carrying one non-identifying property. That is the whole list, and it is the whole of what we ask for:
- receipt_open — which entity’s receipts were opened, and which page it was opened from
- outbound_vendor — the hostname of the vendor site you left for
- faq_open — which page, and the first 60 characters of the question you expanded
- prompt_copy — which guide and step a prompt was copied from — never the text you copied
- dataset_download — which of the three open-data files was downloaded — the filename, nothing else
Nothing that identifies a person is sent with any of them, and nothing is stored on your device to link one click to the next. The aggregate counts live in Vercel’s dashboard, visible to us as totals; the reporting window is set by our Vercel plan, and the visitor session itself is discarded after 24 hours.
Google Analytics, added 20 August 2026
On 20 August 2026 we added Google Analytics 4 (property G-44CBEZQND1) alongside the two Vercel tools. We added it because the question this site cannot currently answer is whether anybody is finding it from a search engine at all, and Google’s own tool is the one that reports that against Google’s own search data.
This one is not cookieless, and that is a real change. Google Analytics sets first-party cookies in your browser — _ga and _ga_G-44CBEZQND1 — which Google documents as being used to distinguish one visitor from another and to persist session state. Before this date the site set no cookie on readers at all, and earlier versions of this page said so. That sentence is no longer true, so it has been removed rather than left standing.
What it records is page addresses, referrer, approximate location derived from your IP address, device and browser type, and how long a visit lasts. We have not enabled Google Signals, advertising personalisation, remarketing audiences, or any user-provided-data matching, so nothing here is used to follow you to another site or to advertise to you. Google acts as our processor for this data and may process it outside the UK under its own transfer safeguards.
How to opt out, in order of least effort. Any tracker blocker or browser with tracking protection switched on will stop it, and so will Google’s own opt-out browser add-on ↗. Clearing your cookies for this site removes the two named above. If you would rather we deleted what is already recorded against your visits, email us; we can act on a request under the UK GDPR whether or not we can identify you from it.
Cookies: the full list
There are three, and this is all of them.
_ga and _ga_G-44CBEZQND1 are set by Google Analytics when you read a page, as described in the section above. They are first-party cookies, they carry a randomly generated identifier and not anything about you, and they are the only storage this site places on your device as a reader.
aiintel_admin is a signed session cookie set only when the site’s own editor signs in to the private admin area. If you are reading a ranking, it is never set.
Vercel Web Analytics and Vercel Speed Insights, the other two tools described above, set nothing at all.
Serving the page: what our host sees
Delivering any web page requires the server to receive the request, and that request contains your IP address. This site is hosted by Vercel Inc., which acts as our processor under its own Data Processing Addendum. Vercel’s privacy policy states that when an end user interacts with a customer’s site it collects “End User IP address, location information derived from IP address, and system configuration information” — the ordinary mechanics of serving a page and keeping the platform up.
We do not build profiles from that, we do not analyse it, and we do not join it to anything else. Vercel is a United States company, so serving this page involves a transfer of that request data outside the UK, under the terms of Vercel’s addendum. Vercel’s privacy policy ↗
What is deliberately absent
- No third-party fonts or assets. Fonts are downloaded at build time and served from this site. Next.js documents this plainly: “No requests are sent to Google by the browser.” Every image, script and stylesheet is served from this domain.
- No advertising technology. No ad network, no demand-side platform, no retargeting tag, no conversion pixel, no consent-management vendor.
- No behavioural tooling. No session recording, no heatmaps, no A/B testing tool, no chat widget, no push notifications, no device fingerprinting.
- No social embeds. No like buttons, no embedded posts, no video players that phone home while you read.
- No newsletter, and no list. We do not hold an email list, so there is nothing to unsubscribe from and nothing to leak.
What leaves this site, and where it goes
Our own URLs, to search engines. When new pages are published we submit the list of their addresses to IndexNow, the protocol Bing, Yandex, Seznam and Naver use, so a new ranking is discoverable the hour it ships. That submission contains our URLs and nothing else — no reader is involved in it.
Aggregate search data, from Google to us. Google Search Console shows us which search queries this site appeared for and how many people clicked. That is Google’s data about its own users, aggregated by Google before we see it; it contains no individual and we cannot identify anyone from it.
You, when you follow a link out. When you click through to a vendor’s site you are on their property under their privacy policy, and their server sees your request the same way any server would. Where an outbound link earns us a commission it is labelled (paid link) on its face, and any tracking of that click happens at the destination, set by the vendor’s network, never by us. How we make money
Nothing, from our public API. The read-only API at /api/v1 serves the same published facts the pages do. It needs no key, no account and no personal data.
If you email us
Then we have your email address and whatever you put in the message, because that is how email works. We use it to answer you and, where you have reported an error, to fix it and record the correction. We do not add you to any list, because there is no list.
Correspondence is kept while the matter is open and afterwards only where it forms part of the record of a published correction. A correction itself stays published permanently — that is the point of it — but it names you only if you have asked to be credited. Ask us to delete your correspondence and we will.
Your rights under UK GDPR
Lawful basis. We do not ask for consent to store things on your device because we do not store anything on it. To the extent the aggregate page-view measurement described above is personal data at all, our lawful basis is legitimate interests — UK GDPR Article 6(1)(f) — namely knowing which pages are read so we can decide what to test next. The interest is narrow, the data is not identifying, and there is no profiling. For email we received from you, the basis is the same, or simply doing the thing you asked us to do.
What you can ask for. A copy of any personal data we hold about you, its correction, its erasure, restriction of its use, or that we stop processing it. In practice the only personal data we are likely to hold is an email you sent us, so those requests are quick to answer. Write to corrections@aiintelligence.com and we will respond within one month, as the regulation requires.
Opting out of the measurement. Any content blocker that blocks analytics scripts stops the page-view beacon, and nothing on this site depends on it. Every page is static HTML that reads perfectly with JavaScript switched off — the only things that need scripts are the drop-down menus, and the footer lists every page on the site without them.
Complaints. If you think we have handled your data badly, tell us first and we will fix it. You also have the right to complain to the UK’s regulator, the Information Commissioner’s Office, at ico.org.uk ↗.
Children. This site is written for adults choosing software they pay for. It is not directed at children and we knowingly collect nothing from them — which is easy, because we collect nothing from anyone.
If this changes, this page changes first. Adding anything that watches a reader — an ad network, a newsletter, a cookie — means rewriting this page in the same release, and dating it. Read next: how we make money, the terms of use, or how to reach us.