The whole documented corpus, as one dated file you can hold

Every lookalike storefront hostname we have documented and published — 61 of them across 57 brands, carrying 653 dated observations between them — exported as one table, one row per hostname, fixed to the day you buy it. The rows are the ones re-checked in our re-verification round of 27 October 2026; each carries the state we observed in our fixed vocabulary, the date we observed it, every superseded observation with its own date, and the artefacts we published for it, quoted verbatim.

Of the 61 hostnames on this index: 34 were live, HTTP 200 when we last checked them; 21 do not resolve; 3 resolve but do not serve; 3 we could not re-check, because our own check failed. A hostname's state can change on any day, in either direction — one row on this list was observed “does not resolve” on 31 July 2026 and “live, HTTP 200” again on 2 September 2026, re-confirmed 8 September 2026, 15 September 2026, 22 September 2026, 13 October 2026, 18 October 2026, 19 October 2026, 20 October 2026 and 27 October 2026.

What arrives, exactly

• ·A 16-column CSV, one row per documented hostname — 62 lines including the header row, covering all 61 hostnames across 57 brands.
• ·The full dated observation history: all 653 observations we hold, every superseded reading kept with its own date in the row it belongs to, oldest first.
• ·Per row: the brand the page reproduced, the evidence class it falls in, the state we observed in the fixed three-string vocabulary, the date of that observation, the date it was last observed serving, and the date it was first documented.
• ·The artefacts we published for each hostname, quoted verbatim — filenames, asset paths, product codes and reproduced copy — plus the flag for the .shop clone-kit build and a link to that hostname's own public record.
• ·The annotations that keep the dates honest: the rows where our own check failed (the date does not advance, and the row says so) and the rows re-classified when what they served changed (the earlier class is kept with its own date).
• ·The same table as a paginated PDF, 24 pages, with the dated summary and the limits on the front of it — the copy you can circulate inside a team or attach to a file.

The 16 columns are, in the file’s own order: hostname, brand, evidence_class, observed_state, observed_on, observed_on_iso, last_observed_serving_on, first_documented_iso, decay_history, not_rechecked_note, reclassified_note, artefact_count, artefacts, clone_kit_domain, record_url, dispute_page_url.

Every row is a dated record of what we observed on the day we checked it, in the past tense, and never a statement about what any hostname is serving today. A hostname that stopped resolving is not described as removed, suspended, seized or taken down: we did not observe why it stopped, and we read back no registry hold for it.

An excerpt from the round of 27 October 2026

6 rows from the round of 27 October 2026, exactly as the file carries them. 58 of the 61 rows completed a reading on that round; 3 did not, because our own check failed, and those keep both their earlier finding and their earlier date. Nothing below is redacted — every row is already published free on its own page, and the link in each row goes to it.

6 rows of 61. The file carries all of them, and the column the excerpt cannot show on a phone is the one worth most: every superseded observation for that hostname, each with its own date, in the row it belongs to.

How to check any row above for yourself

Every reading in the file was taken this way, and you can re-run any of them yourself against any hostname in the excerpt. A check you run today will not necessarily agree with ours: ours is dated, yours is today's, and the difference between the two dates is the finding rather than an error in either.

dig +short <hostname> A
curl -s 'https://dns.google/resolve?name=<hostname>&type=A'

curl -sS -o /dev/null -w '%{http_code}\n' -H 'Cache-Control: no-cache' https://<hostname>/

curl -s -H 'accept: application/rdap+json' https://rdap.gmoregistry.net/rdap/domain/<name>.shop
curl -s -H 'accept: application/rdap+json' https://rdap.verisign.com/com/v1/domain/<name>.com

Decay is not one-way, and the file is built around that

The finding this corpus exists to carry: one row on this list was observed “does not resolve” on 31 July 2026 and “live, HTTP 200” again on 2 September 2026, re-confirmed 8 September 2026, 15 September 2026, 22 September 2026, 13 October 2026, 18 October 2026, 19 October 2026, 20 October 2026 and 27 October 2026. That is why every row keeps its whole observation chain rather than a single current state, and why no row is ever written as finished. A hostname that is not in the zone today can be back in it tomorrow, and one row here has already done exactly that.

We did not observe why it stopped resolving, we make no claim that it was removed, suspended or taken down, and a hostname that stops resolving can be re-pointed at any time — one in this corpus already has.

Read the file before you buy it

Five real rows in full, the complete column schema with a worked example for every column, and the CSV and PDF exactly as they are delivered — produced by the same builders the paid export runs, so nothing here can read better than what arrives.

What $499 does not buy

• ·You are not buying permission. The published dataset is free and stays free under the Creative Commons Attribution 4.0 International (CC BY 4.0), including commercial use, with attribution as the single condition — the rows, the method and a machine-readable release are all at /lookalike-domain-dataset, and nothing on this page is withheld from it.
• ·You are not buying exclusivity, and nobody loses access because you bought it. The same corpus stays published, and the same /lookalike-domain-dataset.csv stays downloadable by anyone.
• ·You are buying the artefact and its delivery: the whole corpus in the evidence-class shape, as a loadable CSV and a paginated PDF, fixed to the day of purchase, with a receipt, an order page and an invoice route for a procurement record.

The free releases are the same records in another encoding: the dataset page, CSV, JSON and plain text, under CC BY 4.0 (Creative Commons Attribution 4.0 International (CC BY 4.0)). If that is enough for what you are doing, take it and do not buy this.

Not included, in plain words

• ·No monitoring. We do not watch these hostnames for you after delivery, and nothing in the file is re-checked on your behalf.
• ·No feed and no API. There is no endpoint, no webhook, no push and no alerting — this is one file, bought once.
• ·No ongoing updates. The export is fixed to the day it was produced. A later re-verification round does not update your copy, and buying this does not entitle you to the next one.
• ·No registrar, hosting provider, network, name server or IP, for any hostname. That attribution is withheld here exactly as it is on every free page of this site.
• ·No takedown, no filing, no outcome and no timing of any kind — this is a record of what we observed, not an action taken against anyone.

The scope, in our own register’s words

These are not sentences written for this page. They are the scope limits from our own claims register, reproduced word for word, and they bind every surface of this site — including the one you are reading.

• ·Every report is a point-in-time audit of publicly observable information, dated to the day the reads were taken. A hostname's state can change at any time after that date.
• ·No continuous monitoring, on any tier. There is no real-time feed, no API and no alerting, and nothing is re-checked after the report is delivered.
• ·No takedown execution, on any tier: we never file. We assemble the evidence and draft the request; the buyer sends it.
• ·We never contact the registrant, and we do not contact a registrar, host or platform on a customer's behalf.
• ·No outcome and no timeline is promised. Whether anything is suspended is the registrar's, host's or platform's decision alone.
• ·No legal representation, ever. This is the boundary that did NOT move and must never be softened. We are not a law firm.
• ·No fund recovery.
• ·No legal proceedings, and no advice on the merits of a trademark position.
• ·We make no assertion about any registrant's intent or criminality

Quoted from growth/approved-claims.md §13b, with no word added, removed or softened — and it binds this page as much as any other surface here.

We describe what each page serves; we make no assertion about any registrant’s intent or criminality, and nothing here is legal advice.

If this is the wrong shape

Want one brand’s rows rather than all 57? Every brand in this corpus has its own slice on the dataset page, free, with that brand’s full dated chains on it. Want the index as a reader rather than as data? The documented index lists every hostname by the brand it reproduces. Want your own domain swept instead of our corpus read? That is a different report — start with the free count at /brand-audit, which asks for no email, no card and no account. A question first: scam-detective8@mail.acoco.ai.