Skip to main content
Use the LangChain MCP adapter to load Atlan tools, then pass the resulting tools to the agent or graph node that needs registry context.
Preview. The connection is provided by langchain-mcp-adapters; it is not a native Atlan package or a direct LangGraph configuration format.

​
Prerequisites

  • A Python project using LangGraph or LangChain.
  • langchain-mcp-adapters added through your package manager with a committed lockfile.
  • The Atlan MCP endpoint: https://agentgateway.atlan.engineering/mcp.
  • OAuth support in the runtime.
  • Workspace access to the tools the graph should use.

​
Setup

  1. Add the Atlan endpoint to MultiServerMCPClient with Streamable HTTP.
  2. Configure the runtime’s OAuth sign-in flow for the Atlan connection.
  3. Load the discovered tools and pass only the required tools to the graph node or agent.
Do not place credentials in graph source or endpoint URLs.

​
Expected result

MultiServerMCPClient returns the Atlan tools available to the authenticated identity as LangChain-compatible tools.

​
Verification

Confirm that the client lists the expected tool names. Then execute one read-only tool through the graph and verify its result before enabling change-capable tools.

​
Limitations

  • This integration is in Preview and is not enabled for every Atlan account.
  • The adapter is stateless by default and creates a fresh MCP session for each tool invocation.
  • Tool errors can be returned to the model as tool messages; connection and authentication failures still raise.
  • The integration inherits LangChain and LangGraph lifecycle and model costs.
See the LangChain MCP guide for sessions, authentication, tool filtering, and custom configuration.