Search with MCP
Connect your AI tools directly to your Atlan catalog using the Atlan MCP server and the Model Context Protocol (MCP). Ask questions in Claude, Cursor, ChatGPT, or any other MCP-compatible tool, and it queries Atlan in the background to return real-time asset metadata—definitions, owners, certification status, lineage, and more—without leaving your workflow. This is the discovery interface for developers and data engineers who work directly in AI tools.
Supported tools
Connect your preferred AI tool to Atlan using one of the guides below:
Interactive AI assistants
| Tool | Setup guide |
|---|---|
| Claude (browser + Claude Desktop) | Set up → |
| Cursor | Set up → |
| ChatGPT | Set up → |
| Gemini | Set up → |
| VS Code | Set up → |
| Windsurf | Set up → |
Automation and workflow tools
| Tool | Setup guide |
|---|---|
| Microsoft Copilot Studio | Set up → |
| n8n | Set up → |
Remote MCP is enabled for all Atlan tenants. No infrastructure setup is required—authentication and permissions follow the same policies you already use in Atlan.
Example discovery prompts
Once your AI tool is connected, use prompts like these to find and explore assets:
Find all tables related to revenue, billing, or invoicing across all our systems. Show which ones are verified and who owns them.
Find the top 10 high-popularity tables related to customers that are certified.
Find frequently accessed but uncertified tables from the Snowflake connection.
For the table `orders_summary`, list all downstream assets and identify impacted dashboards.
Get all downstream assets for the table CUSTOMER_TRANSACTIONS.
Find the 'riskiest' tables: tables in the top 25% by popularity score that are missing two or more of: owner, description, certification, and lineage.
These prompts work across all supported MCP-connected tools. For the full library of prompts—including asset management, lineage, governance, and data quality—see Use MCP in chat-based tools.
Next steps
- Atlan MCP overview: Full list of what the MCP server can do beyond discovery
- Remote MCP setup: How Remote MCP works and supported connectors
- Local MCP setup: Run the MCP server locally with Docker or uv