World-Model MCP for Finance Ergen has launched a beta MCP server that exposes its financial relationship database as callable tools — entity lookup, relationship traversal, pathfinding, StreetQuery and event feeds — usable by any agent that speaks MCP, including Claude, GPT, Gemini and Kimi, with no custom integration. The server connects at https://app.ergen.ai/mcp and is configured via a .mcp.json file, while every Ergen object is also reachable through authenticated REST; the company says a pathfind query returning 3 paths in 0.6s is typical. Ergen is in alpha and reviewing access requests from investment firms, analysts, risk and compliance teams, and research firms. Risk moves through relationships and interactions. We quantify and index them, helping you and your agents build better models. Between any two entities, we record every action, statement, transaction, and event — typed and scored. You can answer questions no terminal can touch. See how exposure travels through suppliers, investors, governments, executives, intermediaries, and events. Via direct query or a browsing experience on our UI. Go from abstract schema to real-world entity sets matching it. These questions would take a team of researchers a week, and now its just a query. Asking complex open-ended questions is not penalized by a huge cost anymore. Harness engineered for financial research and risk analysis, backed by native access to Ergen's DBs. Explore how remote theoretical events could indirectly affect your portfolio, and be ready for it. Our MCP server exposes every Ergen capability as a callable tool — entity lookup, relationship traversal, pathfinding, StreetQuery, event feeds. Any agent that speaks MCP can use it directly. No custom integration required. Works with Claude, GPT, Gemini, Kimi , and any other LLM with tool-use support. // .mcp.json — drop this in your project root { "mcpServers": { "ergen": { "type": "http", "url": "https://app.ergen.ai/mcp" } } } Every object in Ergen is accessible via authenticated REST. Build portfolio monitoring tools, integrate intelligence into your workflow, or feed structured news data into your own models. MCP server in beta — use it from any AI agent. // GET /v1/pathfind · 3 paths · 0.6s { "chain": "R. Mehra", "P. Sharma", "A. Sharma Min. Finance " , "rel types": "family", "professional" , "confidence": 0.94 }, { "chain": "D. Vo", "Aisha Rahman", "I. Chowdhury Jt. Sec. " , "rel types": "family" , "confidence": 0.88 } Your real exposure isn’t the position. It’s the web of relationships, interactions, and events around it. Currently in alpha. Reviewing access requests from investment firms, analysts, risk/compliance teams, and research firms.