Patlytics Patent Search

Concept-level, AI-native patent intelligence — inside Claude.

Describe an invention in plain language and get semantically similar patents, look up a specific patent's claims, search academic literature, and browse your organization's private patent portfolios — all from a Claude conversation.

Who can use it

The connector is for existing Patlytics users. Each member signs in individually with their own Patlytics account, and results are scoped to their organization. Don't have an account? Visit patlytics.ai.

Requirements

Connect

Claude.ai / Claude Desktop: Settings → Connectors → Add custom connector → enter the URL below → Connect, then sign in with your Patlytics account.

https://mcp.patlytics.ai/mcp

Claude Code (CLI):

claude mcp add --transport http patlytics https://mcp.patlytics.ai/mcp
# then in Claude Code:  /mcp  → select "patlytics" → Authenticate

The first time you use a tool you'll complete a one-time sign-in; after that your session persists and refreshes automatically.

What it can do

ToolWhat it does & parameters
Find similar patentsFind patents semantically similar to a described technical concept. Returns concept-level semantic matches over patent claims — useful for prior-art discovery or mapping the landscape around an idea or a reference patent.
  • concept_description — A detailed, prose description of the invention or technical concept. This search is semantic (not keyword-based), so a rich description of how the invention works, the problem it solves, and its key components yields the most relevant matches.
  • reference_patent_number (optional) — Optional. A publication number (e.g. US1234567B2) to anchor the search around a known reference patent.
Get patent detailsFetch a specific patent's bibliographic data and claims by publication number.
  • publication_number — The patent publication number to look up, e.g. US1234567B2 or EP1234567A1. Include the country prefix and (if known) the kind code.
Search non-patent literatureSearch non-patent literature (papers, journals, books) for a topic or keywords.
  • query — A topic, concept, or set of keywords to search across non-patent literature — academic papers, journals, conference proceedings, and books (via OpenAlex). Rich, specific phrasing yields the most relevant results.
List your Patlytics portfoliosList the patent portfolios (workspaces) in your Patlytics organization. Returns your org-private portfolios, each with an id you can pass to list_vault_patents to see the patents inside it.
List patents in your Patlytics vaultList the org-private patents saved in your Patlytics vault. Covers patents your organization has saved or uploaded (including unpublished drafts), optionally scoped to a single portfolio.
  • portfolio_id (optional) — Optional. Restrict to the patents in one portfolio (workspace), by the id returned from list_portfolios. Omit to list the whole organization vault.

All tools are read-only. Deeper analyses — freedom-to-operate, infringement, invalidity, and drafting — run in the Patlytics web app, linked from every result.

Example prompts

Tip: search is semantic, not keyword-based — richer descriptions (how the invention works, the problem it solves, its key components) return more relevant matches.

Privacy & data

The connector authenticates you via Patlytics' identity provider and accesses your Patlytics account and organization data to serve results; non-patent-literature searches query OpenAlex. See the Patlytics Privacy Policy.

Support & troubleshooting