Patent & Trademark RAG

Convert patent and trademark documents into RAG-ready Markdown chunks. Built for IP counsel, brand teams, researchers and AI agents.

⏰ Run this as an IP corpus builder (recommended)

Feed it publication numbers (or Google Patents URLs) and it returns embed-ready Markdown chunks:

{
  "items": ["US10000000B2", "US9876543B2"],
  "chunkWords": 800
}

Weekly cron (0 8 * * 1) turns this into an always-on IP corpus builder.

📊 What you get

Clean JSON, one record per Markdown chunk (6 fields):

• source — the publication number or URL you supplied
• url — the Google Patents page the text came from
• title — document title
• chunkIndex — 0-based position of this chunk in the document
• totalChunks — number of chunks the document was split into
• markdown — the chunk text, RAG-ready Markdown

Pricing: $0.04 per chunk (Pay-Per-Event) — about 25 chunks per $1.

🤖 Use with AI agents

Point Claude, the OpenAI Agents SDK, an n8n flow or any MCP-aware client at it and run IP documents converted to RAG-ready chunks on demand.

Sample agent prompt:

Convert a patent into clean Markdown chunks I can embed for retrieval.

Agentic payments (x402): Supports agentic payment via x402 — agents can call this actor with USDC, no API key required.

🔗 Related IP actors

Clear a mark across every major registry: USPTO Trademarks, EUIPO Trademarks, WIPO PatentScope, CNIPA China, Korea KIPO

Prior-art & patent web: USPTO Patent Search, USPTO Patent Grants, USPTO Grants Tracker, WIPO PatentScope, PatentsView Inventors

IP agent front door: Patents & Trademarks IP MCP

---

Source: Google Patents (patents.google.com) full-text publication pages.

Search globally first: Google Patents Scraper — one worldwide prior-art search across every jurisdiction.