Almost a year ago I built a small proof of concept: talk to patent offices directly through an MCP server. Then the daily usage started climbing and never stopped. Agentic AI is what tipped it over, and that little experiment quietly became something people now run every day. Today I
Patents are written in a secret language: the kind you get when legalese and engineering specs have a child and let a filing cabinet name it. That is how you end up with classification codes like C02F1/4604, which sounds like a malfunctioning robot but actually means "removing salt
Among patent offices, the USPTO runs one of the most ambitious open-data programs out there. The Open Data Portal (ODP) consolidates what used to be a scatter of legacy hosts into a single, key-authenticated, documented surface: patents, assignments, continuity, office actions, PTAB, bulk datasets. There is a real OpenAPI spec,
After a tough review process, OpenAI just published Patent Connector to the ChatGPT app directory - and I couldn't be more thrilled. Anyone can now add it with a single click, no developer mode required, and with OpenAI's seal of approval that the app is safe
I added trademark image retrieval to Patent Connector. The MCP specification defines exactly how a server should return images to a client. I implemented it by the book. Then I tested it across clients and the results were all over the place. What the Spec Says The MCP specification defines
Patent Connector version 1.2.0 adds USPTO office action search and trademark lookup tools. These pull data from the new ODP endpoints that USPTO migrated from the legacy Developer Hub. The underlying Go client library was updated to version 1.4 to support these APIs. Office Action Tools You
Version 1.4 of the uspto-odp Go client library adds two major API categories: Office Action search and Trademark Status & Document Retrieval (TSDR). USPTO migrated both from the legacy Developer Hub into the Open Data Portal, and the library now covers all of them. As with previous releases, I
Bulk File Loader reached version v1.0.1 with a new data source: the German Patent and Trademark Office (DPMA). The application now downloads bulk patent data from DPMA alongside existing sources, built on top of the dpma-connect-plus Go client library I released earlier. What It Does Bulk File Loader
My consulting philosophy: never talk about something you don't fully understand and know. So when MCP came up as a topic, I built a patent MCP server to get real implementation experience. I also found the concept behind MCP very interesting - a standardized protocol for giving LLMs
I've released a Go client library for the German Patent and Trademark Office's Connect Plus API. The library provides access to all three DPMA services - patents and utility models, designs, and trademarks - through a clean, production-ready interface. Technical Approach No official OpenAPI specification exists
Thomson Reuters down as much as 18%. RELX down 14%. Wolters Kluwer down 13%. Not because patent attorneys or contract lawyers became obsolete, but because Anthropic showed that the software wrapping their workflows can be replicated with an open-source plugin. The expertise is fine. What's dead is the
Starting 2026 with a fresh release: https://github.com/patent-dev/bulk-file-loader is now open source. Bulk File Loader automates downloading patent bulk data from EPO and USPTO. Set it up once, configure your credentials, and let it handle the rest - scheduled syncs, progress tracking, and webhook notifications included. Features: