Trademark Search
Exact lookup and similarity screening parameters, pagination, billing and error codes for POST /v1/trademarks/search.
One endpoint covers both exact lookup and similarity screening: POST /v1/trademarks/search. A search costs 10 credits per page; similarity scoring costs another 10 credits per item.
Basic search
curl https://api.lagrange.so/v1/trademarks/search \
-H "Authorization: Bearer $LGR_KEY" \
-d '{ "q": "NOVA", "offices": ["US"], "limit": 20 }'Similarity screening (clearance scenario)
Adding similarity enables five parallel screening channels, and every hit carries an explanation:
{
"q": "KOMM",
"offices": ["US", "EM"],
"similarity": ["identical", "fuzzy", "embedded", "phonetic"]
}{
"results": [
{ "mark_text": "KOMMI", "similarity": { "tier": "near_identical", "via": "fuzzy", "score": 0.94 } },
{ "mark_text": "KOMMEL", "similarity": { "tier": "similar", "via": "fuzzy", "score": 0.81 } }
]
}tier: risk level (identical>near_identical>similar>weak)via: the channel that matched, explaining why the record was returned- Long shared prefixes are weighted:
LAGRANGEvsLAGRANGIANhas edit distance 3, yet still scoresnear_identical
Filter parameters
| Parameter | Type | Description |
|---|---|---|
q | string | Mark name or applicant keyword |
offices | string[] | Jurisdiction filter; defaults to all |
nice_classes | int[] | Nice classification |
status.stage | string | live / pending / dead |
similarity | string[] | Which similarity channels to enable |
limit / page | int | Up to 100 per page, default 20 |
Enums are not exhaustive here
Full jurisdiction codes, status enums and the event dictionary come from the reference endpoints (0 credits): /v1/offices, /v1/status-codes, /v1/event-types.
Known limitations
- The lookalike (visual similarity) channel is in private beta and not enabled by default
- A single search window is capped at 10,000 rows; add filters for deep pagination
- No translation:
mark_textkeeps the original language, so search using the original text