[pd]PatentDistance

REVERSE SEARCH · [PD] CORPUS

Search the claims, not the marketing copy.

Paste what you are building. We break it into limitations and score each one against every claim limitation in the corpus, so you can screen your design before you take it to a patent attorney. It is a semantic prior art search: matches are ranked by meaning, not keywords.

SEARCHING 18,943,997 CLAIM LIMITATIONS · 351 GRANT WEEKS (2020 → 26) · SOFTWARE CPC CLASSES ONLY

Software patents only today: data processing (G06F), AI (G06N), business methods (G06Q), graphics (G06T), networking (H04L/H04W). Every patent lookup in the app feeds this corpus. Submitted designs are never stored.

18.9M+
claim limitations
469K+
granted patents
351
grant weeks
2020→26
grant window

HOW A SEARCH RUNS

01

You paste a design

Plain English, two or three sentences. "A wearable band that unlocks a door when the wearer's phone is in range."

02

We split and embed

Your description is broken into limitations, each embedded and matched against every claim limitation in the corpus by meaning, not keywords.

03

You get the closest art

Ranked claims with match scores, so you can see which element of your design is the close one. Screen any result in one click.

RUNS WITH THE REST OF THE SITE

Reverse search

This page. The box up top searches the whole corpus by meaning.

Run a search ↑

Clearance check

Full claim-by-claim distance score against one patent, with the same engine that powers screening.

Run a check

Design-around brainstorm

After a check, get engineering alternatives built to break the risky limitations, each scored by our engine.

How it works

Questions about the corpus

How complete is the coverage?
Deep in software categories and growing weekly from the USPTO grant feed. It is not yet exhaustive of every software grant, and it holds no other categories today. Use it to find the closest candidates fast, not to certify that nothing else exists.
Why does a description work better than keywords?
The search is meaning-based. Describe the mechanisms: what the parts are, how they connect, what happens to data. That is what the embeddings score.
Is my description stored?
No. It is embedded, searched, and discarded. Only public patent text lives in the corpus.
What does the score mean?
Cosine similarity between your description and a claim limitation, 0 to 1. It ranks candidates. It is not a legal opinion on infringement.
How is this different from keyword patent search?
Keyword search matches words; patent claims rarely use engineering words. A claim about "a sealing collar that radially compresses against a rotating surface" describes a shaft seal, but never says "shaft seal". This search embeds meaning: your description and every claim limitation become vectors, and results are ranked by how close the meanings sit.
Where does the corpus data come from?
The USPTO’s own bulk grant full-text files, parsed weekly as new patents grant. Everything here is public data: granted patent text only. Your search description is embedded, searched, and discarded.
What is a claim limitation?
One clause of a patent claim: a single element the invention is required to have. Claims are scored limitation by limitation because that is how infringement is actually tested, and it shows you exactly which part of your design sits closest to existing patents.
Can I search by patent number?
Yes. Type a patent number like US9589543B2 (or just 9589543) into the search box and this page shows that patent’s full claims, with a link to screen it against your design. To find close prior art from a description of your own design, use the text search instead.