Patent search Boolean operators: building a search string

Step 1: split the invention into concepts

Write one sentence describing what is new, then pull out the two to four ideas that must all be present for a document to matter. More than four and almost nothing passes; one and you are browsing. The running example in this guide is a method for recalibrating a vehicle's lidar while the vehicle is driving. Three concepts: the sensor, the calibration, the setting.

Step 2: expand each concept with synonyms

Patent drafters avoid everyday words on purpose, and different companies settle on different terms of art. The USPTO's own search strategy starts here: an umbrella could also be a parasol or a sunshade.

Where to find the words: the claims and abstracts of the first relevant documents you find, the titles of the CPC groups that cover the field, acronyms and their expansions, US and British spellings (Patent Public Search has a British Equivalents switch for this), and broader generic terms, since claims often say "optical sensor" where the description says lidar. Keep the list and add to it every time a relevant document uses a word you had not tried.

Step 3: join with Boolean operators

OR the synonyms inside each concept, then AND the concepts: (A1 OR A2 OR A3) AND (B1 OR B2) AND (C1 OR C2). Two habits save a lot of grief.

• Always use brackets. Espacenet applies Boolean operators from left to right, and so does Google Patents, whose help gives the example that safety OR seat belt is searched as (safety OR seat) AND belt.
• Know the default. Two words with nothing between them mean AND in Espacenet, Google Patents and The Lens. In Patent Public Search they take the Default Operator from the search options, which the USPTO's training material describes as ADJ, so they may be read as a phrase.
• Be wary of NOT. NOT radar also throws away the lidar patent that mentions radar once in its background section. Prefer narrowing with fields or proximity.

Step 4: tighten with proximity and fields

In a 40-page specification, almost any two common words appear somewhere. Proximity fixes that by requiring the concepts to be discussed together: lidar NEAR8 calibrat$ means within eight words. Field limits do the same job differently: a match in the claims means the feature is part of what was protected, not a passing mention. A reasonable progression is full text with sentence-level proximity first, then title, abstract and claims with looser proximity, then claims only.

The big caveat is Google Patents. Its help says proximity operators "only change ranking, not retrieval", so a NEAR query returns the same documents as AND, ordered differently. For proximity that actually filters, use Patent Public Search, Espacenet or The Lens.

Step 5: widen with truncation, carefully

Truncation turns calibrat into calibrate, calibrated, calibrating, calibration and calibrator in one term. The traps are database specific:

• Patent Public Search will not run a query if one term matches more than 20,000 terms, or if the truncation numbers multiply to more than 125 (FAQ).
• Espacenet needs three characters before *, caps ? and # at three or seven per word, has no left truncation, and allows no wildcards on classification symbols (pocket guide).
• Google Patents ORs only the 25 most common matches of a wildcard, on single English words, so a short stem silently loses rare forms (help).
• The Lens does not stem wildcard terms, so with stemming switched on a term like valve* may not match as expected; turn stemming off in Structured Search (search syntax).

Short stems are the usual mistake. car* finds carbon and cartoon. Four or five letters is a sensible minimum.

Step 6: anchor it with a CPC class

Keywords only find documents that use your words. A classification finds documents examiners put in the same technical box, whatever words they use. For the lidar example, CPC G01S 7/497 covers means for monitoring or calibrating lidar-type systems. Two patterns work well together: a narrow class with loose keywords (class AND concept B), and a broad class with tight keywords (subclass AND concepts A, B and C in proximity). Our CPC classification guide covers how to find and read the right class.

Syntax compared across four databases

Each row is taken from the provider's own documentation: the USPTO's Advanced Search guide and searchable indexes, the EPO's Espacenet pocket guide, Google Patents Help and The Lens's search syntax and classification pages. "Not documented" means the provider's help does not describe it, not that it cannot be done.

Two more differences worth knowing. Espacenet's Advanced search form does not accept AND, OR or NOT inside a field; use the field operators any and all instead (EPO note). And The Lens indexes classification only at the group a document is filed in, not its parents, which is why its CPC search needs a trailing wildcard and an escaped slash.

The same search in each syntax

The lidar example, concepts A and B in the claims (or title, abstract and claims) and concept C anywhere:

1. Patent Public Search

   ((lidar OR (laser ADJ scanner)) NEAR8 (calibrat$ OR recalibrat$ OR misalign$)).clm. AND (vehicle OR car OR automobile)

   Proximity filters. Turn on Plurals in the options to catch vehicles and cars.

2. Espacenet

   (ctxt=lidar OR ctxt="laser scanner") AND (ctxt=calibrat* OR ctxt=recalibrat* OR ctxt=misalign*) AND ftxt any "vehicle car automobile"

   Concepts ANDed in title, abstract or claims. To tie two words together, add a proximity clause such as ctxt=(lidar prox/distance<8 calibrat*).

3. Google Patents

   (lidar OR "laser scanner") AND CL=(calibration OR recalibration OR misalignment) AND (vehicle OR car OR automobile)

   Plurals and close synonyms are added automatically. A NEAR operator here would reorder results, not filter them.

4. The Lens

   claim:((lidar OR "laser scanner") AND (calibrat* OR recalibrat* OR misalign*)) AND full_text:(vehicle OR car OR automobile)

   Turn stemming off in Structured Search if you rely on wildcards. For proximity, use a quoted pair with ~n, such as "lidar calibration"~8.

Build these in stages, not in one go. Run each concept as its own set, note the counts, and combine. Patent Public Search saves each statement as an L number you can reuse (L1 AND L2); in the other tools, keep the strings in a text file. Either way, record the database, date and exact string of every search, as the USPTO's strategy recommends. Database-specific walkthroughs: Patent Public Search, Espacenet and Google Patents.

Semantic search vs keyword search

Keyword search returns documents that contain your terms in the pattern you specified. Semantic search compares meaning: you describe the invention and it returns documents about something similar, even if they share few words with your description. The official tools are keyword tools. The USPTO says plainly that Patent Public Search "does not use semantic searching". Google Patents sits in between, adding synonyms automatically and offering a list of similar documents on each result.

The weakness of semantic search is that it is hard to audit: you cannot easily say what it did not look at, and a ranked list has no natural end. The weakness of keyword search is that it is only as good as your synonym list. Used in sequence they cover each other.

More patent guides

• Patent drawing requirements: USPTO, EPO and India

  37 CFR 1.84, the EPO rules that replaced Rule 46 EPC, and Indian rule 15, compared side by side.

• How to write patent claims: USPTO, EPO and India

  Claim structure and worked examples, then the claim rules and fees at the USPTO, the EPO and the Indian Patent Office, side by side.

• Types of patent searches, and which one you need

  Patentability, freedom-to-operate, invalidity and landscape searches compared by purpose, timing, scope and deliverable.

• Patentability search: how to check novelty before you file

  What counts as prior art under 35 U.S.C. 102 and Art. 54 EPC, the grace period gap, search steps and how to read results.

• Freedom-to-operate search in the US and Europe

  In-force claims only, legal status checks at the USPTO, EPO and UPC, FTO opinions after Halo, and design-arounds.

• Invalidity search: prior art to challenge a patent

  IPR, PGR, EPO opposition and UPC revocation deadlines, prior art dates, standards documents and claim charts.

• Patent landscape analysis, step by step

  Building a CPC and keyword dataset, cleaning families and names, the metrics that matter and the 18-month publication lag.

• Google Patents search guide

  Search operators and fields, Boolean and proximity syntax, non-patent literature, and where Google Patents falls short of the official databases.

• The Cooperative Patent Classification (CPC), explained

  How CPC symbols are built, CPC against IPC, the Y section, how to find the right class and how to search it in each database.

• The PCT application: timeline, fees and national phase

  The international phase month by month, receiving Office and ISA fees in CHF, USD and EUR, national phase deadlines by office, and a deadline calculator.

• How much does a patent cost?

  USPTO fees by entity size from filing to the last maintenance fee, what drives professional fees, EPO fees, and an estimator.

• How long does a patent last?

  US utility and design terms, patent term adjustment and extension, terminal disclaimers, maintenance fee deadlines, Europe and India, with an expiry estimator.

• Can AI write a patent application?

  What AI drafting does well and badly, the USPTO and EPO positions on AI tools, AI inventorship after 2025, and confidentiality.

All guides, by patent office