Housecarl
AI assistants

Connect Arbiter to your AI assistant with the Model Context Protocol (MCP)

The Model Context Protocol (MCP) is an open standard that lets an AI assistant call outside tools during a conversation. An assistant that supports MCP can connect to Arbiter, send it the sources and claims for a question, and get back a verdict with the reasoning attached.

You do not need to write code for this. In Claude you paste one web address and sign in; the Quick Start has the steps for each assistant.

The pages below explain what the connection gives you, and when the REST API, the plain programming interface for software you write yourself, is the better choice. For the work each role does with it, see Arbiter for your work.

  1. An evidence-weighing Model Context Protocol (MCP) server for Claude, ChatGPT and Cursor

    What Arbiter adds to an AI assistant over the Model Context Protocol: named sources, reliability weights, visible contradictions, and a verdict you can audit.

  2. Model Context Protocol (MCP) or REST API: which to use

    Arbiter has two doors: the Model Context Protocol for AI assistants, and a REST API for your own software. Which one fits, and what each one needs.