Atlan MCP: Enterprise Context Layer

Atlan's MCP server brings the full power of the Atlan data catalog into any AI assistant or agentic workflow. It enables natural language interaction with your metadata, allowing users to discover assets, trace lineage, manage governance artifacts, and maintain data quality without leaving their AI tool of choice. The listing covers core capabilities across discovery, governance, asset management, and data quality, validated through a structured suite of positive and negative test cases.

**Use cases**

* Intelligent data discovery: search for tables, columns, dashboards, and other assets using natural language queries with contextual metadata, descriptions, and relationships.
* Lineage and impact analysis: trace upstream sources and downstream dependencies to understand data flow and assess change impact before deployment.
* Data governance: create and manage business glossaries, terms, categories, domains, and data products to support data mesh architectures.
* Asset management: update descriptions, READMEs, certifications (VERIFIED, DRAFT, DEPRECATED), and attach glossary terms or custom metadata to improve discoverability.
* Data quality: create, schedule, update, and delete data quality rules to monitor catalog health continuously.

**Product details**

Sample test cases and triggered tools represent the breadth of metadata operations supported by the listing.

* Datasets represented include Customer Data Glossary, Sales & Marketing domain, customer\_dim table, DIM\_CUSTOMER table, and FCT\_SALES table.
* Sample fields include glossary name, description, certificate status, asset GUID, and qualified name.
* Tools triggered across positive test cases include create\_glossaries, create\_domains, semantic\_search\_tool, update\_assets\_tool, and traverse\_lineage\_tool.
* Positive test scenarios cover glossary creation, domain creation, semantic search, asset updates, and lineage traversal.
* Negative test scenarios validate scope boundaries: SQL data queries, physical table DDL, external file system searches, and database permission management are correctly excluded as out of scope.

For more details, refer to the embedded notebook.

**Additional Insights**

The listing has been validated against a structured QA suite combining positive cases (where specific tools must trigger) and negative cases (where the MCP must correctly defer or decline). Negative cases confirm that the server distinguishes catalog metadata operations from adjacent concerns such as transactional data queries, physical schema management, file system search, and access control administration. This separation of concerns ensures predictable agent behavior and reduces the risk of unintended actions when the MCP is composed with other tools in a broader AI workflow.

**To Properly Login**
Login Instructions
1. Click on Login. A web page will open in your browser.
2. Enter your Atlan domain and sign in to your Atlan workspace.
3. Once login is successful, Databricks will receive the credentials automatically.
4. The MCP will be set up and ready to use.

If you have any issues or questions, please contact -
Gene Arnold - [e-mail]