Frequently asked

ClaimHit does not rely on a single model's opinion. Several models propose candidates in parallel, and every candidate is then cross-checked against live web evidence. Only a product confirmed on public sources reaches your results, with those sources listed so you can open them yourself. Anything that could not be confirmed is dropped rather than shown with a caveat. Every result is a lead worth investigating; none is a determination of infringement.

US (USPTO), European (EPO), WIPO PCT, and most national offices via Espacenet. Enter any standard format, such as US10123456B2, EP1234567B1, WO2020123456A1. Japanese, Korean, and Chinese patents are supported via EPO translation.

A ClaimHit search is fast and broad, ideal for initial screening across many potential infringers, with every lead cross-verified against live web evidence. Expert Review is a manual review of your AI Hit Chart by a subject matter expert, who checks each claim element and fills gaps against verified evidence. Use a search first to identify promising targets, then order Expert Review for those.

A result is a company and product that survived the full pipeline: proposed by several models, checked by live web search, filtered for noise and category, and confirmed on the manufacturer's own page. It is listed with the pages, datasheets or filings that confirmed it. Open the sources first; a result you have read is a lead, a result you have only looked at is a name.

Because infringement is a legal conclusion, not a research output. ClaimHit surfaces candidates that may infringe: products and standards worth investigating, each cross-verified against live web evidence. Determining whether infringement actually exists requires a qualified patent attorney and formal claim construction. Our role is to make that investigation faster, not to replace it.

Several frontier models from different providers run in parallel against your patent's inventive contribution. Each independently proposes candidate products and companies. This is the discovery step, not the final answer. Using models from different providers matters: they have different training data and failure modes, so a candidate several of them surface independently is less likely to be one model's hallucination. Every candidate is then cross-checked against live web evidence and confirmed on manufacturer pages before it reaches you. Models propose; evidence decides.

Yes. Your searches are used only to generate your results and are never sold or used to train AI models. Patent numbers are public records. We treat any additional context you provide as confidential.

Standards Essential Patent analysis checks whether your patent covers mandatory features in technical standards like Wi-Fi 6, 5G NR, or HEVC. SEP holders can demand FRAND (fair, reasonable, non-discriminatory) licensing from any implementer of the standard. ClaimHit maps your claims to published specifications from 3GPP, IEEE, ETSI, ITU-T, and IETF.

One patent: the full product search and one Hit Chart, for a fixed price shown on the homepage, with a 30-day money-back guarantee. Once per account, no subscription, no card on file after that. It is the fastest way to judge ClaimHit on a patent you know inside out.

Ask for a credit pack from the Credits page in your account or through the contact page. Tell us how many product searches and Hit Charts you need and we reply with a proposal within 24 hours on a business day. Credits never expire.

Yes. Paste the list on the homepage form. We reply with a proposal for the whole list within 24 hours on a business day and run it on your go-ahead: the Hit Matrix across the portfolio, claim charts on the strongest hits, and the campaign briefs. No software to learn.

A few minutes for the search, and a few more for a Hit Chart on the strongest hit. Results arrive on screen and by email.

A claim chart built element by element. The claim is broken into its limitations, and each is judged FULL, PARTIAL or NONE against a quotable excerpt from public evidence: a datasheet, a manual, a filing, a standard. A rating only survives if a real excerpt supports it. Nothing is rounded up.

When you run a portfolio, the results come back as one grid: patents along the top, companies down the side, a filled cell where a patent reaches one of that company's products. The companies that keep appearing across patents are where a licensing conversation starts. The grid exports to Excel.

A free door. Give us a patent number and a work email, confirm the address, and we read the companies whose products reach the patent, look again every quarter, and email you how many new ones appeared. The names and the sources are the Hit Report, which the pilot runs on any patent. Up to three patents per address, for a year, one click to stop.