Ask 30+ knowledge graphs from an AI assistant

WOBD can be used through guided forms, but the broader federation is best explored conversationally. Connect Claude, ChatGPT, or any MCP-capable client to a single Model Context Protocol endpoint. The assistant can discover relevant graphs, inspect schemas, bridge identifiers, run graph queries, and synthesize results while preserving the source graphs behind the answer.

Built on the mcp-proto-okn project; see the unified server documentation for the full tool list and design. Connect your AI assistant →

1. Connect

Point Claude, ChatGPT, VS Code, or another MCP-capable client at the hosted Proto-OKN endpoint.

2. Explore graphs

The assistant can discover relevant graphs, inspect schemas, bridge identifiers, and run graph queries.

3. Audit answers

Ask which graphs were used and request the query trace so synthesis stays grounded in source records.

What to ask

  • Find datasets that support a disease, exposure, gene, pathway, or drug question.
  • Compare evidence across graphs that normally live in separate portals.
  • Ask which source graphs were used and request the queries behind an answer.

Why it matters

General chat can invent connections. WOBD-backed assistant workflows ground answers in queryable graphs, shared identifiers, and inspectable source records, making exploratory synthesis easier to audit.

Example vignettes

Connect Claude Desktop or ChatGPT

Use the public endpoint and a short verification prompt to confirm the assistant can see the Proto-OKN graph tools.

More worked analyses across the Proto-OKN federation

Browse the upstream mcp-proto-okn example library — spanning environmental justice, criminal-justice patterns, spaceflight biology, drug-target discovery, and more.