ipnex.ai

IPNEX serves cited, confidence-stamped patent intelligence over US university technology-transfer (TTO) out-licensing: each marketed technology joined to its real patent picture -- current assignee, live legal status, expected expiry. Call 'search_licensable_ip' to find currently-licensable IP (filter by institution, status, keyword, min_confidence, and licensable_from_institution); call 'get_technology' for the full published view of one TTO docket (point-in-time via 'as_of'). Every result carries an 'ownership' signal -- whether the patent's current assignee of record is the marketing institution or a third party (the patent may not be licensable from that university); set licensable_from_institution=true to exclude the third-party cases. Every result also carries a 'status' interpretation -- the honest reading of the legal status, distinguishing a term-expired patent (public_domain=true: in the public domain, free to practise, nothing left to license) from a revivable fee lapse or a still in-force patent -- and an 'expiry' reading: how many years of exclusive term remain (years_remaining) as of a reference date, and whether it is expired / expiring_soon (under ~2 years left) / active_term, so you can tell a patent marketed as live but nearly out of term from one with two decades to run (status says whether a right exists; expiry says how long it lasts). Each result also carries a 'licensability' score: a gated 0-100 read of whether it is actually licensable, composed from those same ownership + in-force + remaining-term facts (each component shown openly; indicative when an input is unknown). A search response also carries 'facets' -- corpus-level breakdowns over all matches by ownership (institution-owned vs assigned-away / third-party / unknown), by institution, by status (in force vs lapsed vs other vs unknown), by status_detail (the finer split: term-expired vs fee-lapsed vs reinstated ...), and by expiry (expired / expiring_soon / active_term / unknown) -- so you can read the whole-corpus split for a query without paging every result. Every field carries its source citation and a confidence score; unresolved fields are omitted, never guessed -- an unknown docket returns {"found": false}, not an error. Patent data: Google Patents (CC BY 4.0) + USPTO (public domain). Summaries are IPNEX-authored. TTO listing content is linked, not reproduced.