Skip to main content

MCP Integration Overview

Knowledge Tree provides a Model Context Protocol (MCP) server that gives any MCP-compatible AI client read-only access to the knowledge graph. This means you can explore nodes, facts, dimensions, and relationships directly from tools like Claude Desktop.

What is MCP?

The Model Context Protocol is an open standard that lets AI assistants connect to external data sources and tools. Instead of copying data into prompts, MCP provides structured, on-demand access to live data.

What the Knowledge Tree MCP server provides

The MCP server exposes 8 tools for navigating the knowledge graph:

ToolPurpose
search_graphFind nodes by text search
get_nodeLoad node details, definition, counts
get_dimensionsLoad multi-model analyses (paginated)
get_edgesLoad relationships, sorted by evidence strength
get_factsLoad facts grouped by source, with powerful filtering
get_fact_sourcesLoad deduplicated source list for provenance
search_factsSearch the global fact pool
get_node_pathsFind shortest paths between two nodes

All tools are read-only — they query the graph but never modify it.

Authentication

The MCP server uses Bearer token authentication. You generate an API token from the Research App's profile page, then include it in your MCP client configuration.

Getting started

  1. Connect your MCP client
  2. Explore the available tools
  3. Try common workflows