How we detect what a site is running

Ad networks are identified from public advertising records — the standardised files and tags that publishers and networks maintain so that advertisers can verify who is behind an ad slot. None of it is a guess about how a site looks, and none of it appears unless someone deliberately set it up.

No listing here is a paid placement, and none is accepted by submission. We don't take money to add, move, or remove a site.

1. ads.txt records

ads.txt is a file publishers host at the root of their domain to declare which companies are authorised to sell their ad inventory. It exists specifically so that buyers can verify authorisation, and it is public by design: it has to be openly readable for the system it belongs to to work.

When a site's ads.txt contains a network's declared reseller lines, that is a direct, publisher-authored statement that the network is authorised to sell for them. This is our strongest signal.

2. Tag script signatures

Networks load ads through a script served from a domain they control — for example scripts.mediavine.com/tags/, g.ezoic.net, or pagead2.googlesyndication.com with a ca-pub- publisher ID. Finding that script in a page confirms the network is live on the site rather than merely authorised.

3. Verification and freshness

Our index is compiled from public ad-configuration data and is being progressively re-verified against live sites by our own crawler. Rows carry the date they were last confirmed, so you can see the age of what you are looking at rather than taking a freshness claim on trust.

We would rather say plainly that a large index is verified on a rolling basis than imply every one of tens of thousands of rows was checked this morning. When a site stops matching, it is marked removed rather than deleted — a list that silently drops entries cannot be audited, and knowing that a publisher left a network is often more useful than knowing they joined.

4. What we deliberately do not collect

We do not collect or publish personal contact details, email addresses, or anything scraped from a site's pages. The "Contact" link on each row is simply the conventional /contact path on that domain, constructed rather than harvested. Publisher names are the business or brand names sites publish about themselves.

5. Limitations, stated plainly

• A site can authorise a network in ads.txt without actively running it, and can run one without a matching record. We weight both signals and mark the basis for each row.
• Detection reflects a site's homepage. Networks used only on subsections may be missed.
• Sites behind aggressive bot protection may be unverifiable, and are excluded rather than guessed at.
• Our coverage is a large sample, not a census of the entire web. Counts should be read as "at least this many".
• Verification is rolling, not simultaneous. A publisher who changed networks very recently may not be reflected yet.

Corrections

If your site is listed incorrectly, or you would like it removed, contact us and we will amend the record.