What is Patalyze

Patalyze is the context layer you plug into your AI agent to analyze large amounts of patents and related data. It contains tools to fetch, organize and present patent data in your own research databases.

The data tools access our global corpus of patents, commercial products parsed into structured feature trees, and mappings that map these records to another with claim charts. The organization tools let you package data into isolated research databases and extend data records with your own attributes. The presentation tools create pages like tables, dashboards, and notes, that turn those data records into something a person can understand and trust.

Research databases#

Everything in Patalyze happens inside a research database, a self-contained home for one analysis. It holds the patents you gather, the products you track, the mappings between them, and the pages you build to review the results. Each database is fully isolated, so data never crosses from one into another.

A new database never starts from zero. The patents you add are drawn from a shared global corpus, and the products from a shared catalog; what makes a database yours is the selection and the analysis you build on top.

As a rule, one database answers one question. Clearing a single product for launch is one; scanning the market for products that might infringe a patent family is another. Reusing an existing database saves the setup, but a verdict is only as trustworthy as its scope, and a dedicated database keeps each conclusion easy to stand behind months later. Name it for the question it answers, and when related databases pile up, group them in a folder.

Data#

Every analysis is built on four kinds of data. Patents are the rights you search and pull in, by the million if you need to. Products are real-world things turned into structured features. Mappings read one product against one patent's claims. Attributes are the fields you add to extend the records with data of your own.

Presentation#

Pages are how you read and present a research database. A table narrows rows down to the dangerous few, a dashboard charts the shape of the whole dataset, and a note writes up what you found. Every page is one of these three.

Workflows#

Data and presentation compose into the analyses themselves. Each is a path through the same pieces, in a different order.

Underneath, both follow the same mechanism. Add patents and products to a research database, and Patalyze creates a mapping for every pair, scoring each claim, element by element, against the product's features. The results land on pages you can read: tables to filter and sort, dashboards to chart the findings, notes to write them up. What changes between workflows is the question and the order you read in, not the machinery.

Integrations#

Two MCP servers connect Patalyze to AI clients, so an agent can run an analysis the same way you would. Both authenticate over OAuth: add a URL, approve access in the browser, and there is no API key to manage.

Setup takes a few minutes. The flow is the same in every client; what differs is where you add the server URLs. Pick yours for step-by-step instructions:

Prefer to talk to Patalyze directly? The Data API runs boolean and semantic search across millions of patents over plain HTTP, returning full bibliographic records and patent-family data, with every field and operator documented in the Endpoints reference.