Map your endpoints

The MarkerAPI column shows the operation segment after /api/v2/trademarks/. See the MarkerAPI v2 documentation and the Markbase API reference for request parameters.

MarkerAPI v2MarkbaseBehavior to check
serialnumberGET /trademark/{serialNumber}Lookup by USPTO application serial number. The Markbase response is a single record.
trademarkGET /search?q=…Fuzzy search across the mark, owner, and goods and services. Exact and wildcard queries need a different approach.
descriptionGET /search?q=… (partial)The same full-text search includes goods and services. There is no description-only substring endpoint.
ownerGET /owner?name=…Searches the owner field. MarkerAPI wildcard patterns need to be replaced with plain search terms.
expiringNo equivalentMarkbase has no expiration-window search. Keep this workflow on its current data source.

Change the base URL and authentication

Use https://api.markbase.co. MarkerAPI puts its username and password in the URL path. Markbase takes a separate key in the X-API-Key header. Keep it in server-side secret configuration as MARKBASE_API_KEY. Remove the old credential segments from the URLs your client builds.

curl -fsS -G 'https://api.markbase.co/search' \
  --data-urlencode 'q=nike' \
  --data 'page=1' \
  --data 'limit=10'

The request above uses keyless access, which has lower limits. For a keyed integration, create a Free account and copy the key from your dashboard.

Update result parsing and pagination

Read search and owner results from hits. A serial-number lookup returns the record directly. For common fields, map serial number to serial_number, mark text to word_mark, goods and services to goods_services, and status code to status_code. Keep serial numbers as strings and handle nullable fields. For classification, read international_codes or us_codes according to the classification system your client needs.

curl -fsS 'https://api.markbase.co/trademark/90792662'

curl -fsS -G 'https://api.markbase.co/owner' \
  --data-urlencode 'name=Nike, Inc.' \
  --data 'page=1' \
  --data 'limit=10'

Replace MarkerAPI's next and start flow with page=1 and an explicit limit. Keep your query and limit fixed as you increment page. The API caps the limit by plan, so compare hits.length with the returned limit to detect a short page. Set a maximum number of pages and treat total_hits as an estimate.

Recheck exact matches and statuses

Send plain text in q or name. Markbase does not implement MarkerAPI's asterisk syntax. If you need exact word-mark equality, compare the returned word_mark values using your application's normalization. Filtering a bounded set of fuzzy hits can miss records, so it cannot reproduce a complete exact or substring search.

For a MarkerAPI query using status=all, omit the Markbase status filter. MarkerAPI documents active as an option but does not list its underlying codes. Markbase's status_code filters by one USPTO code, so define which statuses your application needs and check that rule against known records. If you filter statuses after receiving a page, use its unfiltered hit count to decide whether to page again.

Handle errors before switching traffic

Treat 404 on a serial-number lookup as a missing record and an empty hits array on a successful search as zero results. Check your key if Markbase responds with 401. Both a per-minute limit and a monthly quota return 429. If the response includes X-Quota-Reset, defer until the next quota window. For a per-minute limit, use X-RateLimit-Reset to time a bounded retry. Both reset headers use Unix seconds.

Compare known serial numbers, owner spellings, and pending, registered, and inactive records with your existing integration. MarkerAPI's public v2 docs do not specify its error responses, so test those with your existing client. Use search results for preliminary screening. An empty result cannot establish that a name is legally available.

Keep reading