Patents MCP
MCP server for searching Google Patents on BigQuery for prior art and inventor lookups
- Pricing
- Open source
- Platforms
- CLI, API, macOS
- Version
- 1.0.0
- Listing updated
- today
Similar MCP servers
About Patents MCP
Patents MCP is an open-source MCP server for legal and government. It is available on CLI, API, macOS, Windows and Linux.
The README describes patents-mcp as a Model Context Protocol server that lets AI clients run patent searches against the Google Patents public dataset hosted on BigQuery. Instead of relying on a commercial patent database, it queries that open dataset directly through the user's own Google Cloud project.
The server exposes three tools: keyword and inventor search, fetching full details for a publication number, and retrieving the claims text of a patent. It is written in Go using the official Go MCP SDK, and the README notes a configuration that takes a service account key and a GCP project ID through environment variables.
It suits developers and researchers who already use an MCP-capable assistant such as Claude Desktop and want it to help with prior art discovery. Setup requires a Google Cloud project with the BigQuery API enabled and a service account, so it is less suited to users who want a no-setup patent tool.
Tools this server provides (3)
Taken from the project documentation; the running server may differ.
- search_patents
- Search patents by keyword, technology area, or inventor name
- get_patent
- Fetch full patent details by publication number
- get_patent_claims
- Fetch patent claims text
What it does
- Keyword and inventor search
- Searches patents by keyword, technology area, or inventor name.
- Full patent details
- Fetches the complete record for a patent given its publication number.
- Claims retrieval
- Returns the claims text of a specified patent.
- BigQuery-backed data
- Queries the Google Patents public dataset on BigQuery through the user's own GCP project.
- Modular tool design
- Each tool is self-contained, and the README describes adding a tool as a new file plus one registration line.
- Structured logging
- Uses zap-based structured logging with an optional debug mode.
Good for
- Finding prior art before filing a patent application
- Checking what patents a company or inventor has filed on a topic
- Reviewing the claims of a specific patent from within an AI assistant
- Exploring patent activity in a technology area during research
Pricing
Open source
Free and MIT licensed; BigQuery usage billed through the user's Google Cloud account
Patents MCP FAQ
Is patents-mcp free to use?
The README says the server is MIT licensed and built on the Google Patents public dataset, which has a free BigQuery monthly allowance of 1TB. Google Cloud use is subject to its own terms.
What do I need to set it up?
The README lists Go 1.21 or later, a Google Cloud account, the BigQuery API enabled on a project, and a service account with the BigQuery Job User and Data Viewer roles. It is configured with GOOGLE_APPLICATION_CREDENTIALS and GCP_PROJECT_ID.
Which AI client can use it?
The README includes a configuration example for Claude Desktop and an example of the server being used through that client.
Is Patents MCP free?
Yes. Patents MCP is open source, so you can use and self-host it for free.
What are the best alternatives to Patents MCP?
The closest alternatives to Patents MCP are WritersLogic legal-ip agents, Priorbit — IP database for agents, Arcangel AI, Patlas — Patent Prior-Art Search and G62 Patent Search.
Where this comes from
- Description: written from the project's README and its MCP Registry entry.