How to Conduct a Patent Search in India Before You File

Most Indian founders find out too late that patent drafting is not where the real risk sits — the risk is filing on an invention that already exists somewhere in the world, in a database nobody checked. A patent search before you draft can save months of legal fees and a rejected application. It is one of the cheapest, highest-leverage steps in the entire process, and it is routinely skipped.

"Prior art" is anything, anywhere in the world, that has already been made public before your filing date — a granted patent, a published patent application that never even proceeded to grant, a research paper, a product manual, a conference poster, a YouTube demo, or a page on a company website. Novelty in Indian patent law is measured against all of this, not just against other patents. A search that only checks granted patents is checking a fraction of what actually counts.

A few free databases cover most of what you need for a first-pass search. The Indian Patent Office's own InPASS portal covers Indian filings. WIPO's PATENTSCOPE and the European Patent Office's Espacenet both index patent applications from dozens of countries and are free to use without an account. Google Patents is the fastest starting point because it also full-text searches non-patent literature and links to citing and cited documents, which is often how you find the real prior art — not through your own keywords, but through what an examiner or another applicant already cited.

Keyword searches alone miss a large share of relevant results, because two documents describing the same invention rarely use the same words. This is where patent classification codes matter. Every patent application is tagged with International Patent Classification (IPC) or Cooperative Patent Classification (CPC) codes that describe the technical field regardless of the vocabulary used. Finding even one closely related document and pulling its classification codes, then browsing everything else filed under those codes, usually surfaces far more relevant prior art than another round of keyword guessing.

It also helps to be clear about which kind of search you actually need, because the two are often confused. A novelty or patentability search asks whether your invention is new enough to file — it looks broadly at what exists. A freedom-to-operate search asks a completely different question: whether making, using or selling your product would infringe someone else's currently in-force patent rights in a specific country. You can have a perfectly novel and patentable invention that still infringes an existing patent if it incorporates a component someone else already owns rights to. Founders preparing to launch, not just to file, usually need both.

The most common mistake we see is stopping at the first page of comfortable-looking results and concluding the field is clear. A thorough search usually needs several search sessions, multiple keyword variations including synonyms and regional spellings, at least one classification-based pass, and a check of non-patent literature such as academic databases and product catalogues. It also needs an honest, adversarial read of the results — actively looking for reasons your invention might not be novel, rather than reasons it might be.

A do-it-yourself search is a genuinely useful first filter, and every founder should run one before spending on anything else. But it should not be the last word before you file. A professional search adds trained classification searches, access to paid databases that index more non-patent literature than the free tools, and — most importantly — an independent read on how close is too close. At Ragulika IP, we run novelty and freedom-to-operate searches as the first step for founders and R&D teams, precisely so the drafting stage starts from a realistic picture of the landscape rather than an assumption that the field is clear.