Expired Domain API: Automating the Search

Checking expired domains by hand works until it does not. Somewhere around fifty candidates a day, the manual pass stops being viable and you start skipping the checks that matter. That is the point to automate — and the trick is knowing which parts to hand over.

What automation is good at

Fetching drop lists. Zone files from ICANN CZDS publish daily and are large. Pulling and diffing them to find newly expiring names is mechanical work a script does better than a person.

Filtering out the obvious rejects. Most of any drop list is noise — no backlinks, no history, machine-generated names. Rules discard the bulk of it without judgement.

Checking link profiles at volume. Pulling referring domains for a few hundred candidates and scoring them against a list of sources you trust is exactly the repetitive comparison worth scripting.

Watching for drop dates. A domain you want releases at a specific time. Tracking that across dozens of candidates is a job for a calendar, not memory.

What it is bad at

Reading history. Whether a site's archive shows a coherent subject or a topic change is a judgement call. You can automate fetching the captures. The reading stays yours.

Assessing fit. Whether a domain suits what you plan to build is not expressible as a filter.

Spotting the unusual. The reason a domain has one surprising link, or why a respectable-looking history has a gap, is where the interesting cases live — and they are precisely the ones rules do not catch.

A sensible split: automation narrows a thousand candidates to twenty. You look at the twenty.

Building the pipeline

Ingest. Pull zone files daily, diff against yesterday, emit new expiries. CZDS access requires an application, which is free but takes a few days.

Enrich. For each candidate, fetch backlink data. This is the step that costs money — a backlink API charges per lookup, so filter hard before you reach it. Discarding names with no history first cuts the bill substantially.

Score. Compare referring domains against your list of trusted sources. Output a small structured record: which recognised sources link to it, whether those links are live, and where they point.

Review. Push the survivors somewhere you will actually look — a daily digest is enough. This is the human step, and it should be short by design.

Rate limits and cost

Two things determine whether this stays practical.

Backlink APIs meter by lookup, and the enrichment step is where a naive pipeline burns its budget. Order your filters so the cheap ones run first.

Zone files are large and publish on a schedule. Pull once daily and cache. Do not re-fetch per query.

Whether to build it

If you buy a few domains a year, the manual process is fine and a pipeline is not worth maintaining.

If you are working through drop lists regularly, the automation pays for itself quickly — not because it finds better domains, but because it stops you skipping checks when the list is long. That is the actual failure mode, and it is the one worth engineering against.

Related: what to look for in a domain tool and the buying process.