Back to Blog

How to Conduct a Patent Novelty Search Before Filing

April 26, 2026 Aruna 13 min read
patent novelty search tool — intellectual property law and protection

Before investing thousands of dollars in a patent application, every inventor should conduct a thorough patent novelty search — also called a patentability search — examining existing patents, published applications and other prior art to determine whether an invention is genuinely new. The hard part is rarely effort; it is knowing which patent novelty search tool answers which question, and where each one stops. This guide sets out the process, the databases and how to read what you find.

A patent novelty search is a systematic review of publicly available prior art to assess whether an invention is novel and non-obvious — the two primary requirements for patentability under U.S. patent law. The search examines patents, published patent applications, scientific literature, product manuals, websites, and any other publicly available information that could anticipate or render obvious your claimed invention. While no search can be 100% comprehensive, a well-conducted patent novelty search dramatically reduces the risk of filing a patent application that will be rejected over prior art the examiner finds during examination.

When to bring in a professional search team. A self-serve database session is enough to rule out the obvious, but it rarely settles the question that matters commercially: whether the invention justifies the five-figure cost of prosecution. Professional searchers work from subscription corpora that index non-English and non-patent literature the free interfaces never surface, and they read classification trees rather than keyword hits — which is where most do-it-yourself searching quietly fails.

If you would rather have that work done and documented than run it yourself, PerspireIP’s prior art search services deliver a written report with every relevant reference identified, a claim-by-claim read on how each one bears on your invention, and a patentability opinion you can put in front of a client or an investor. Scope tiers and turnaround are set out on that page, and you can request a quote before committing.

Why Conduct a Novelty Search Before Filing?

The benefits of conducting a patent novelty search before filing go far beyond simply checking whether your invention already exists.

  • Saves money: discovering that your invention is already patented before filing saves the significant cost of preparing and prosecuting an application.
  • Improves claim drafting: knowing the closest prior art allows your attorney to draft claims narrowly tailored to distinguish your invention.
  • Identifies design-around opportunities: search results reveal the landscape of existing technology and gaps for broader claims.
  • Informs business decisions: if your invention is clearly not patentable, you can redirect resources to other IP strategies.
  • Reveals competitor activity: patent searches often reveal what your competitors are developing, providing valuable competitive intelligence.

Key Databases for Patent Novelty Searches

Several online databases provide access to patent documents and other prior art. The USPTO PatFT and AppFT databases contain all issued US patents and published US patent applications and are free, comprehensive starting points. Google Patents offers a user-friendly interface with powerful search capabilities across US patents, international applications, and patents from dozens of foreign patent offices. Espacenet from the European Patent Office provides access to over 130 million patent documents worldwide with powerful CPC classification search tools. The USPTO Patent Center allows viewing of full prosecution histories of published applications.

Step-by-Step Patent Novelty Search Process

Conducting an effective patent novelty search requires a systematic approach. PerspireIP uses a proven five-step process for client searches.

Step 1: Define the Invention’s Core Features

Before searching, clearly identify the key inventive features you want to protect. Break the invention down into its essential elements. What problem does it solve? What is the technical mechanism? What are the key structural or functional components? This analysis guides your search strategy.

Step 2: Generate Search Terms

Develop a comprehensive list of keywords, synonyms, and technical terms that describe your invention. Patent documents often use technical jargon that differs from common usage, so searching multiple variations is essential. For a mechanical fastener, you might search fastener, connector, coupling, clip, latch, and dozens of other terms.

Step 3: Identify Relevant CPC Classifications

The Cooperative Patent Classification system organizes technology into a hierarchical system of classes and subclasses. Identifying the CPC codes relevant to your invention allows you to search comprehensively within a technology area regardless of the specific keywords used. You can find relevant CPC codes using the CPC Classification Definition database on the USPTO website.

Step 4: Conduct the Search and Review Results

Execute your searches across multiple databases using your keyword and CPC lists. Review the title and abstract of each result first, then examine the full claims and drawings of promising references. Pay particular attention to the claims — a patent only blocks what its claims cover, and a prior art reference only anticipates your invention if it discloses every element of your claimed invention.

Step 5: Analyze and Document Your Findings

Compile the most relevant references and analyze how they relate to your invention. For each close reference, identify which elements of your invention it discloses and which it does not. The gaps represent your potential patentable invention — the features that distinguish your innovation from the prior art.

Interpreting Patent Novelty Search Results

Finding prior art during a patent novelty search does not necessarily mean your invention is not patentable. It means you need to understand exactly what the prior art discloses and how your invention differs. Even if all individual elements of your invention appear in the prior art, your invention may still be patentable if the specific combination or arrangement is not disclosed or suggested by the prior art. This is the non-obviousness requirement — even if individual elements are known, the combination may still be inventive if a skilled person would not have been motivated to combine them in the same way.

Non-Patent Literature Searches

A comprehensive patent novelty search should also include non-patent literature (NPL). Scientific journals, conference proceedings, textbooks, product catalogs, and even websites can constitute prior art if they were publicly available before your filing date. Databases such as Google Scholar, PubMed, IEEE Xplore, and the ACM Digital Library are valuable sources of NPL. Patent examiners routinely search these sources, and finding relevant NPL yourself before filing allows you to draft claims that address it proactively.

Choosing a Patent Novelty Search Tool: What Each Database Does Well

There is no single patent novelty search tool that covers everything, and the practical skill is knowing which one answers which question. The major free databases differ in what they index, how far back their full text runs, and how well they handle non-English documents — and those differences decide whether a search misses the reference that later invalidates the patent.

  • USPTO Patent Public Search — the office’s own interface, which replaced the legacy PatFT and AppFT systems along with the internal EAST and WEST tools. It offers the same full-text US search examiners use, with proximity operators and CPC field searching. Use it when you need to see exactly what a US examiner would see.
  • Espacenet (EPO) — the broadest international coverage, spanning patent documents from a very large number of national offices, with machine translation and family data. This is the tool that catches the foreign-language reference a US-only search will miss.
  • WIPO PATENTSCOPE — the authoritative source for published PCT international applications, plus many national collections. Its cross-lingual expansion feature is genuinely useful for finding equivalents of a concept expressed differently in another language.
  • Google Patents — the fastest place to start, because its semantic matching and full-text indexing of both patents and a large body of non-patent literature surface relevant documents from a plain-language description. It is a poor place to finish, because you cannot easily prove what it did not search.
  • National registers — DEPATISnet for German documents, J-PlatPat for Japanese ones. Worth the extra step in mechanical and materials fields, where German and Japanese prior art is dense.

Choosing a patent novelty search tool is therefore a sequencing decision rather than a single pick. A defensible workflow uses semantic search to build vocabulary, classification search to define the field, and full-text search to work it systematically:

  1. Start in Google Patents with a plain description of the invention, and harvest the vocabulary and CPC codes that keep recurring in the results.
  2. Move to Espacenet and search by CPC classification rather than keyword, so the search no longer depends on the words a drafter happened to choose.
  3. Run the same classification and keyword strategy in Patent Public Search for the US collection, and in PATENTSCOPE for PCT publications.
  4. Check the families of the closest hits. A reference dismissed in one jurisdiction often has a sibling with different claims.
  5. Finish with non-patent literature — Google Scholar, IEEE Xplore, conference proceedings, product manuals, archived product pages.

Two limits are worth stating plainly, because they are where amateur searches go wrong. First, the eighteen-month publication delay: applications are generally published eighteen months from the earliest priority date, so no tool can show you what was filed in the last year and a half. A clean search result never proves the field is clear — it proves nothing has published yet.

Second, coverage of full text is uneven. Older documents and many non-English collections are indexed by bibliographic data and abstract only, so a keyword that appears solely in the body of a 1985 specification will not be found by any keyword search. This is exactly why classification searching is not optional. CPC codes were assigned by examiners who read the document, and they work regardless of the words in it.

A final point on what the results are for. Under 35 U.S.C. 102, a single reference disclosing every element of the claim defeats novelty; under 35 U.S.C. 103, a combination of references can render the invention obvious. Those are different tests, and a novelty search addresses the first far better than the second. Nor is a patentability search a freedom-to-operate search: an invention can be perfectly patentable and still infringe someone else’s unexpired claims, because patentability asks whether your invention is new and obviousness-free, while freedom to operate asks whose rights you might be walking into.

What Legally Counts as Prior Art — and the One-Year Grace Period

A search is only as good as the legal test it is measured against, and that test is narrower and broader than most inventors expect. Under 35 U.S.C. 102(a)(1), a claimed invention is unpatentable if it was “patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date.” The closing phrase does the most work. It is not limited to patents and journal articles.

A conference poster, a trade-show demonstration, a product listing, a thesis shelved in a university library, a YouTube walkthrough or an archived forum post can all qualify. This is why a search confined to patent databases produces false comfort in software, consumer products and mechanical arts, where the disclosure that matters is frequently commercial rather than patent literature.

A second category sits in 35 U.S.C. 102(a)(2): subject matter described in a patent or in a published application that names another inventor and was effectively filed before your filing date. These documents are the reason a clean result can go stale. An application filed eighteen months before yours may not have published when you searched, yet it still counts against you once it does. No database can show you what has not yet been published, so a result should always be read as current to the publication record on the date it was run, not as a permanent finding.

Against that, 35 U.S.C. 102(b)(1) supplies a genuine safe harbour that is routinely misunderstood. A disclosure made one year or less before the effective filing date is not prior art under (a)(1) if it “was made by the inventor or joint inventor or by another who obtained the subject matter disclosed directly or indirectly from the inventor,” or if the subject matter had already been publicly disclosed by the inventor before that disclosure.

Read it carefully: the grace period excuses your own disclosures, not a third party’s. An independent competitor publication thirteen months before you file is fatal; your own demonstration eleven months before you file is not. A related provision, 102(b)(2)(C), removes commonly owned or commonly assigned disclosures from the (a)(2) category, which matters for corporate groups and university consortia filing in the same technical area.

One further point decides how a result should be interpreted. Section 102 asks whether a single reference discloses every element of the claim. Section 103 asks the different and much harder question of whether the claimed invention would have been obvious from a combination of references to a person of ordinary skill in the art. A search that returns no single anticipating document has resolved the first question only. Two or three references that together cover the claim, with a plausible reason to combine them, remain the most common ground of rejection after KSR v. Teleflex.

What Finding Art Obliges You to Do: The Duty of Disclosure

Searching before filing creates an obligation that not searching does not, and it catches applicants by surprise. Under 37 CFR 1.56, everyone substantively involved in preparing or prosecuting a US application owes the Office a duty of candour and good faith, including a duty to disclose information known to be material to patentability. Art you located is art you know about.

The mechanism is the Information Disclosure Statement. References turned up before filing, and anything that surfaces later during prosecution, should be cited to the examiner on an IDS within the windows set by 37 CFR 1.97. There is no equivalent duty at the EPO, although a European examiner may ask about search results for corresponding applications, so applicants filing in both systems should not assume one practice transfers.

Applicants sometimes conclude from this that searching is a liability and that it is safer not to look. That reasoning is backwards on both halves. The duty attaches to what you actually know, so deliberately staying ignorant does not neutralise the art; it simply means you discover it after spending the filing and prosecution fees, or worse, after asserting the patent.

The downside risk is also concentrated on concealment rather than on disclosure. Withholding known material art can support an inequitable-conduct defence that renders the entire patent unenforceable, and following the en banc decision in Therasense v. Becton Dickinson that defence requires proof of but-for materiality together with a specific intent to deceive. Disclosing a reference costs you an IDS form. Hiding one puts the patent itself at stake.

The practical discipline is to keep a record as you go: the queries run, the classification codes and date ranges covered, the databases consulted, and the references reviewed and set aside with a short note on why. That log is what makes an IDS quick to assemble, lets a later attorney re-run and extend the work rather than repeat it, and demonstrates good faith if the question is ever raised years afterwards.

Professional Search Services from PerspireIP

While do-it-yourself searches using free databases are a good starting point, a professional patent novelty search conducted by PerspireIP’s experienced search team provides a significantly more comprehensive and reliable result. Our searchers have access to professional-grade databases, deep familiarity with CPC classification systems, and years of experience interpreting search results in the context of patent law. We provide a detailed search report with all relevant references identified, an analysis of how each reference relates to your invention, and a preliminary assessment of patentability.

Conclusion

A thorough patent novelty search is one of the smartest investments you can make before filing a patent application. It informs your patent strategy, improves your claims, reduces prosecution costs, and helps you make better business decisions about where to invest your IP resources. PerspireIP is here to help you conduct comprehensive novelty searches and translate the results into a winning patent strategy. Contact us today to get started.