> ## Documentation Index
> Fetch the complete documentation index at: https://platform.atlan.com/llms.txt
> Use this file to discover all available pages before exploring further.

> ## Agent Instructions
> To act on Atlan objects, use the Atlan MCP server at https://api.atlan.com/mcp or the SDK. To install or publish skills from a terminal, use the `aireg` CLI; `aireg --map json` prints its command map. The `atlanai` CLI is deprecated. Run a read-only identity check before any write.
> The docs MCP server at /mcp searches these docs only. It cannot read or change Atlan objects.
> SDK packages: Python `atlanai` (PyPI) and TypeScript `@atlanai/sdk` (npm). Show Python first, then TypeScript.
> These docs cover Atlan AI Registry, Atlan's hosted product. The AI registry specification and aireg CLI are maintained separately. /open-source lists publicly reachable signed installers and explains that the source repository and hosted CLI reference currently require GitHub access. Do not claim the repository or its reference docs are publicly accessible. Use these docs for Atlan-specific setup such as the adapter, sign-in and workspaces.

# Connect Databricks agents to Atlan

> Register Atlan MCP Gateway as an external MCP service in Databricks Unity Gateway and test it in AI Playground.

Register Atlan MCP Gateway in Unity Gateway so Databricks agents can search
Registry artifacts and call the tools their Atlan identity can access.

## Before you begin

Use a Unity Catalog workspace with Model Serving support. Your role needs
`USE CATALOG`, `USE SCHEMA`, `CREATE SERVICE`, and permission to create or use
the HTTP connection. The workspace must reach `api.atlan.com` and
`auth.atlan.com`.

Choose your Atlan identity: a [registered agent](/api/tutorials/register-agent)
for a shared connection, or each person's account. Keep an agent's
`client_secret` in your secret manager. The
[data platform overview](/integrations/how-tos/data-platforms) lists the shared
endpoint and OAuth values.

## Register the external MCP service

1. In Databricks, open **Unity Gateway → MCPs → + MCP**. Select **Connect an
   existing MCP server**, choose a catalog and schema, and name it `atlan`.
2. Create or select an HTTP connection to `https://api.atlan.com/mcp`. If the
   form separates URL fields, use host `https://api.atlan.com`, port `443`,
   and base path `/mcp`.
3. Choose **OAuth Machine-to-Machine** for an Atlan agent, using its client
   credentials, token endpoint `https://auth.atlan.com/oauth/token`, and
   scopes `contextplane:read contextplane:write`. For each person's account,
   choose **Dynamic Client Registration**.
4. Select **Create & load tools** and sign in if prompted. Include
   `mcp__initialise` and the Registry or remote tools the agent needs, then
   select **Create MCP Service**.
5. Grant callers `EXECUTE` on the service and `USE CATALOG` and `USE SCHEMA`
   on its parents.

See Databricks' [external MCP server guide](https://docs.databricks.com/aws/en/ai-gateway/register-mcp-service)
for connection permissions and platform availability.

## Verify in AI Playground

Choose a model with tool support. Under **Tools → + Add tool → MCP Servers**,
choose **External MCP servers** and select `atlan`. Ask it to find an Atlan
skill by name. Confirm a `registry__search_artifacts` call and check Atlan's
**Tool calls** for the session activity. Session records depend on the audit
write succeeding.

If a tool is missing, check the service's tool selection and the Atlan
identity's workspace access.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.