Model Context Protocol

Patent prosecution AI
inside your AI assistant

Connect Abigail to Claude Desktop, Cursor, or any MCP-compatible tool. Query your docket, get examiner intelligence, and trigger response workflows without leaving your AI assistant.

Get started freeView documentation
claude_desktop_config.json
{
  "mcpServers": {
    "abigail": {
      "url": "https://api.abigail.app/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

How to connect

01

Create an Abigail account

Sign up at abigail.app. Add your first patent application to your docket by entering the application number.

02

Generate an MCP API key

Go to Settings in your Abigail account and generate an MCP API key. Keep it secure.

03

Add to your MCP client

Add the Abigail MCP server to your claude_desktop_config.json or Cursor MCP settings. Paste the server URL and your API key.

04

Ask about your docket

Ask your AI assistant: "What Office Actions are due this month?" or "Analyze the 102 rejection in application 17/123,456." Abigail answers with real data from your docket.

Available MCP tools

get_docket

List all pending Office Actions with deadlines and examiner details

lookup_application

Retrieve prosecution history and case details by application number

get_examiner_intel

Get allowance rates, interview patterns, and art unit statistics for an examiner

get_claim_analysis

Retrieve element-level claim analysis and prior art mapping for an OA

get_deadlines

List upcoming response deadlines across your entire docket

start_agent_workflow

Trigger the 10-expert pipeline to begin drafting a response

get_pending_reviews

Get responses that are drafted and awaiting attorney review

get_analytics_summary

Portfolio-level statistics on allowance rates and response outcomes

Frequently asked questions

Ready to connect?

Create a free account, add your first case, and generate your MCP API key in under two minutes.

Create free account