Overview
What is Ontology MCP Server?
Ontology MCP Server ingests ISO 20022 XML banking messages, extracts structured financial signals, builds an RDF knowledge graph, and exposes it as native tools for any MCP-compatible AI agent. It is designed for banks, fintechs, and payment processors who need to query interconnected payment messages.
How to use Ontology MCP Server?
Install via pip install ontology-mcp and run python mcp_server.py. Alternatively, use Docker: docker pull ghcr.io/mrcholis/ontology-mcp:latest and docker run .... Connect any MCP client (e.g., Claude Desktop) by adding the server to your claude_desktop_config.json. A demo can be run with python -m ontology_mcp.demo.
Key features of Ontology MCP Server
- Ingests ISO 20022 XML into a queryable knowledge graph.
- Exposes 11 MCP tools for AI agents.
- Supports pacs.008, pacs.009, camt.053/054/056/026/029.
- Persistent storage with Oxigraph (RocksDB-backed).
- Real-time WebSocket streaming ingest.
- Built-in anomaly detection for high-value transactions.
Use cases of Ontology MCP Server
- Trace the full lifecycle of a payment cancellation.
- Find all transactions over $50
