Document · policy
What this site records when you click a buy link, and what it does not.
FilterSpecs sets no cookies, runs no analytics on its pages, and loads nothing from a third party. Reading this site is not recorded by us at all. The only thing we record is a click on a buy link, and that record contains nothing that identifies you.
Our pages are delivered by Cloudflare, which adds a small set of technical headers to every response. One of them asks your browser to report a failed connection to Cloudflare. Those reports go to Cloudflare, not to us, and we never see an individual report. Every header is listed below.
We do not need to know who you are to run this business, so we have built it so that we cannot. That is a design decision enforced by our build: a page that set a cookie or loaded a third-party script would fail our deploy.
Our pages are static files. They carry no cookies, no local storage, no tracking pixels, no fonts or scripts from another origin, and no analytics of any kind — Google Analytics included. Nothing on a page reports back to us that you read it.
Our hosting provider keeps its own short-lived operational logs, as every web host does, to serve pages and absorb attacks. We do not query them for analytics and they are not part of our dataset.
Cloudflare is our edge provider: it stores our pages and sends them to your browser. Along with each page it sends these headers, and no others. None of them sets a cookie or stores anything on your device except the connection hints your browser keeps for itself.
| Header | What it does |
|---|---|
| date | The time the response was sent. |
| content-type | What kind of file this is, for example a web page or an image. |
| content-length | The size of the file. |
| content-encoding | That the file was compressed for transfer. |
| transfer-encoding | That the file was sent in pieces rather than all at once. |
| cache-control | How long your browser may reuse its copy before asking again. |
| etag | A fingerprint of the file's contents, so your browser can tell whether its copy is current. |
| location | Only on a redirect: the address you are being sent to. |
| server | Names Cloudflare as the server that answered. |
| alt-svc | Tells your browser it may use a newer, faster connection protocol (HTTP/3) next time. |
| cf-cache-status | Whether Cloudflare answered from its cache. Operational only. |
| cf-ray | An identifier Cloudflare gives the request so its engineers can trace a fault. It names the request, not you, and we do not collect it. |
| nel | Network Error Logging. Asks your browser to report a failed connection to Cloudflare. Successful page loads are never reported. |
| report-to | Tells your browser where those failure reports go: a.nel.cloudflare.com, which is Cloudflare, not us. |
| referrer-policy | Only on a buy link's redirect: tells your browser to give the merchant our domain name and nothing more, so the merchant does not learn which page of ours you were reading. |
About the failure reports named by nel and report-to: if your browser cannot load a page from us, it may send Cloudflare a short report naming the address, the kind of error and how long it took. A page that loads successfully is never reported. Per Cloudflare's own documentation, it holds your IP address only while receiving the report, derives from it the country, metro area and network the report came from, and uses the reports to find connectivity problems on its network. The reports are sent to Cloudflare, not to us, and we never see an individual one.
This list is checked against our live responses. A header that is not on it fails that check, so this table cannot quietly fall behind what your browser actually receives.
Buy links pass through our own redirect before sending you to the merchant. That redirect writes one row. Here is every field of it:
| Field | What it holds |
|---|---|
| site | Which of our sites the click came from. |
| product_id | Which product you clicked. |
| merchant_id | Which merchant you were sent to. |
| monetized | Whether that link earned us a commission. |
| referring_platform | One of nine values: chatgpt, perplexity, gemini, claude, google, bing, reddit, direct or other. Worked out from utm_source below, or from the host the request came from when that is not our own site. A click on one of our pages with no tag records direct. |
| referrer_host | The host the request came from. Usually our own site, because buy links sit on our pages. The host only. |
| source_path | Which of our pages the buy link was on, for example /under-sink-ro/apec-ro-90/. The path only. |
| utm_source | The utm_source tag on the address of the page you clicked from, when it had one. For example, ChatGPT adds utm_source=chatgpt.com to the links it gives. Only this one tag is kept. |
| country | The country Cloudflare reports for the request. Country only, never a city or a region. |
| device | One of three values: mobile, tablet or desktop. |
| is_bot | Yes or no: whether the request came from automated software, such as a search engine crawler or a link preview, rather than a person. Decided from your browser's user agent at the moment of the click. Only the yes or no is kept. |
| ts | The time of the click. |
That is the complete list. There is no other field, and the row cannot be tied to you or to any other row.
The page a click came from, and a utm_source tag on that page's address, are facts about the page, not about you. Everyone who clicks buy on the same page after following the same link produces the same two values. We record them to learn which of our pages lead to a purchase and, where an assistant tags the links it gives, which assistant sent the reader. Our buy links ask your browser to tell our redirect the full address of the page they are on. The redirect keeps the two values above, discards the rest, and tells your browser to pass the merchant only our domain name.
If a buy link is broken — the address names no product, or names one we hold no redirect for — we cannot send you to a merchant, so we send you to our own not-found page instead. That writes one row to a separate table, holding the same fields as above except the product and the merchant, plus the address that was asked for, so we can find and fix the link.
It is a separate table rather than a flag on the one above, because a link that leads to no product is not a click on a product and counting it as one would overstate every figure we publish about how this site is used.
Our redirect keeps no log of its own. From 24 to 26 September 2026 it kept short-lived diagnostic logs, to find out why a click had failed to be recorded, and we switched them off on 26 September 2026.
While they were on, each line named a table and whether a write to it succeeded, and Cloudflare's Workers Logs attached the address requested on our redirect, the request method and Cloudflare's own request identifiers. They held no IP address, no user agent, no country and no address of the page you came from. Cloudflare deletes those logs after three days on its free Workers plan and never keeps them longer than seven days on any plan, so the last of them are gone by 3 October 2026 at the latest. We never copied them anywhere or joined them to the rows above.
Because we store no identifier, we cannot build a profile of you, cannot recognise you across visits, and could not hand over a history of your browsing if asked, because no such history exists.
The merchant learns only that you came from FilterSpecs's domain. Our redirect tells your browser not to pass on which page of ours you were reading.
When you follow a buy link you leave this site and arrive at the merchant, who sets their own cookies and applies their own privacy policy. We do not control that and we cannot see it. Affiliate networks may record that a sale originated from us, in order to pay us; they do not tell us who you are, and we do not ask.
We may earn a commission on a purchase. How we make money explains it, and commission never affects our rankings or which products we cover.
Laws including the GDPR and the CCPA give you rights over personal data held about you. We hold none: the click rows contain no identifier, so they are not personal data and there is nothing to export, correct or delete on request.
If you believe we hold something about you, write to hello@specindexgroup.com and we will answer. See contact for how we handle corrections generally.
This site is aimed at adults buying household equipment. We do not knowingly collect anything from children, and since we collect no identifiers from anyone, there is nothing we could collect from a child either.
This policy is versioned and dated at the top of the page. If what we record changes, this page changes in the same release, because it is written from the code that does the recording rather than from an intention.
Aspiros AI, Inc. is the data controller for this site.