Sentence-level patent search · Infringement · Licensing
The world's patents, as vectors.
Every sentence of the full-text corpus, searchable. For R&D teams checking a disclosure, counsel clearing a launch, owners finding who is practising their claims, and platforms building on the layer underneath.
Check a disclosure before you file
Knockout scans over invention disclosures, patentability and prior-art searches at sentence level. File-or-iterate decisions inside the R&D pipeline, with the closest art quoted. The disclosure is never sent to a third-party model to be searched.
R&D and clearance For counsel clearing a launchFreedom to operate
Product features matched against live claims, sentence by sentence. Legal status and family checked. A launch-readiness view that names the risk and the evidence for it.
R&D and clearance For patent owners, counsel and fundersFind who is using your patents
Who is practising your claims, documented to the standard a licensee, a court or a funder expects. Ranked targets, counsel-ready charts, and the prospectus to take them to market.
Infringement and licensing For IP systems and platform teamsBuild on the vectors
Hold the corpus as vectors on your own infrastructure. Encode queries through three APIs. Your language models answer from the record, not from memory.
Platform and methodThe problem
A language model will answer a patent question either way.
What matters is whether it consulted the patent record first.
Invented prior art
Publication numbers that do not exist. Caught on verification. Costly to credibility, not to the decision.
Plausible paraphrase
A real document summarised into something its claims do not say. Survives a spot check and reaches the decision.
Confident silence
The closest analogue was never retrieved, so it is never mentioned. Undetectable. No reviewer knows to look for it.
Better retrieval reduces all three. Only better retrieval can reduce the third.
What we hold
The corpus, as vectors. Searched one sentence at a time.
Engagement
Three ways in. The same retrieval underneath.
Reports
Where most clients start. Knockout, novelty, FTO, competitor monitoring, infringement identification, lapse screening, licensing target lists and portfolio prospectuses. Retrieved evidence drafted through a language model and reviewed by a patent professional. Every value traces to source; unknowns are left blank.
Search APIs
Document and sentence-level searches run on Cintian's systems, called from your own tools. You build the layer above.
Vectors
The corpus and the models, held on your infrastructure. You build your own products and indexes on them.
Retrieval quality is identical across all three. What varies is who is responsible for the layer above.
Pilot
Bring us a disclosure or a patent. We will search it with you.
Ranked results in 20 seconds on average, usually less. A cited report drafted in around five minutes. Measured on your material, not ours.