Domain label search
Finds registrable labels like a search term, and every TLD where each label is listed in the corpus. It is the starting point for typosquatting, impersonation and defensive-registration questions; the investigation that follows is in datapulse_help(topic="lookalike_domains").
The label searched is the PSL+1 label, one level below a public suffix: google in google.com, example in example.co.uk. Hostnames and subdomains are not indexed, so a.b.c.foo.com exists in the corpus only as foo.com (datapulse_help(topic="methodology")). The search matches domain-name labels only: it does not search registration records and cannot find domains by registrant (datapulse_help(topic="rdap")).
Parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
search_term | string | Yes | One label: paypal, not paypal.com. Dots at either end are dropped (.paypal searches paypal); a dot inside is refused. Any text without control, bidirectional-control or zero-width characters, at least the minimum the schema states (3 by default) and at most 255 characters |
filter | array | No | Match types to keep: exact, substring, fuzzy. Omit for all three |
Response
CSV ordered by increasing score, nearest matches first:
label,type,score,tlds
g00gle,fuzzy,2,com|net|org|info
googl3,fuzzy,3,com|net
google-login,substring,6,com|net|org|info|co.uk
Up to two lines can come before the header: error: <text> when the API reported an error alongside its rows, and complete: false (...) when the label limit was reached or the API cut its answer short. Parsing starts at the label,type,score,tlds line. No match is the header alone followed by a hint line (No labels match <term> with <match types>…), prose rather than a CSV row.
| Column | Meaning |
|---|---|
label | The matching PSL+1 label (A-label form for IDNs) |
type | exact, substring or fuzzy |
score | Levenshtein edit distance from the search term: 0 is exact, 1 is one substitution, insertion or deletion away; typical fuzzy results range from 1 to 8 |
tlds | Pipe-delimited list of every TLD where the label is listed: the complete set, never cut |
The number of labels is capped at the limit the tool description states. complete: false (the N-label result limit was reached; more labels likely exist) above the header means more matching labels exist; a narrower filter or a longer term reduces the set.
Listed is not registered. The corpus holds active, functioning names only, so a TLD missing from tlds may still be registered: parked without DNS, not resolving, or too new to be listed. datapulse_live_rdap confirms a gap. A listed label in a TLD does not say who holds the domain; registrant identity is GDPR-redacted and not returned.
A search term containing @ gets a second content item, a WARNING that the tool matches labels and cannot find domains by registrant email. The rows are still returned.
Fuzzy matching
Fuzzy matching is built for lookalikes: homoglyphs (0/o, 1/l/i, rn/m: paypa1, paypai) and IDN confusables (Cyrillic а for Latin a), returned as xn-- labels.
It does not systematically find plain typos (an omitted, transposed or adjacent-key letter). A fuzzy search for paypal (2026-10-06) filled the label limit with IDN variants at distance 1 and left out paypl, which an exact search finds under 18 TLDs; a fuzzy search for a rarer term can return nothing at all. An empty or typo-free fuzzy result is therefore not evidence that typo domains do not exist. Typo coverage comes from generating the variants (omissions, duplications, transpositions, adjacent keys, hyphen changes) and searching each with filter=["exact"]:
datapulse_dns_searchlabels(search_term="paypl", filter=["exact"])
datapulse_dns_searchlabels(search_term="papyal", filter=["exact"])
A fuzzy match listed in 15 or more TLDs is more notable than one listed in a single TLD.
Errors
| Message | Cause |
|---|---|
missing required parameter 'search_term' | No search term |
parameter 'search_term' cannot be empty | Empty or whitespace-only term |
search_term too short: N characters (minimum M) | Below the minimum length |
parameter 'search_term' must be at most 255 characters, got N | Too long |
invalid search_term: ... | A control character (C0, C1, DEL), a bidirectional control (U+202A–202E, U+2066–2069) or a zero-width character (U+200B–200D, U+FEFF) |
search_term is one label (paypal, not paypal.com); "..." holds N (...): search each on its own | A dot inside the term |
invalid filter value "..."; allowed: exact, substring, fuzzy | Unknown match type |
| A timeout | The result set is too large: very common short strings (pay, app) match millions of labels. filter=["exact"] or ["substring"], or a longer term, makes it smaller |
Generated from the live server (DataPulse MCP v0.9-341-g27163c2) on October 20, 2026. Your AI assistant reads this page by calling datapulse_help(topic="searchlabels").