Connect external MCP clients (Claude, Cursor, Replit) to Databricks MCP servers to query data, access Unity Catalog tables and vector indexes. OAuth or PAT; workspace-specific URL.
Databricks connect external mcp clients (claude, cursor, replit) to databricks mcp servers to query data, access unity catalog tables and vector indexes. oauth or pat; workspace-specific url.. The official Databricks MCP server provides AI agents with standardized access to Databricks's services, enabling powerful automation and integration capabilities that enhance developer productivity and streamline workflows.
This official MCP server is maintained and verified by Databricks, ensuring reliability, security, and up-to-date functionality. Unlike community-maintained servers, this official integration provides long-term support, comprehensive documentation, and direct vendor assistance. The server is hosted on databricks.com, guaranteeing secure and reliable access to Databricks's services.
Whether you're building AI-powered applications, automating workflows, or enhancing your development tools, the Databricks MCP server provides a standardized way to connect AI agents with Databricks's extensive platform capabilities. This integration enables natural language interactions with Databricks, making it easier than ever to leverage Databricks's features through AI assistants like Claude, ChatGPT, and Cursor.
https://<workspace-hostname>/api/2.0/mcpGo to Settings → Cursor Settings
Click "Add new global MCP Server"
Paste the MCP configuration into mcp.json and restart
Or manually add to your mcp.json:
{
"mcpServers": {
"databricks": {
"url": "https://<workspace-hostname>/api/2.0/mcp"
}
}
}The official Databricks MCP server enables AI agents to:
These capabilities make it easy for AI agents to work with Databricks as part of a larger automation strategy, enabling seamless integration between AI assistants and your existing tools.
The Databricks MCP server provides 4 tools for AI agents to interact with the platform. Use the search below to find specific tools.
| Tool Name | Description |
|---|---|
get_vector_index | Access vector indexes for RAG |
list_tables | List tables in catalog/schema |
query_catalog | Query Unity Catalog tables and metadata |
run_sql | Execute SQL on Databricks |
Last updated: March 1, 2026 at 12:00 PM
Create, update, and manage tasks in Databricks using natural language. Let AI agents handle project organization and task assignment.
Generate and update Databricks content through AI conversations. Create documents, update records, and manage information effortlessly.
Automate repetitive Databricks workflows. Let AI agents handle routine operations while you focus on high-value work.
Visit https://www.databricks.com, create an account, and generate an API key from your dashboard. Most tools offer a free tier to get started, making it easy to begin using the Databricks MCP server.
Yes, this MCP server is officially provided and maintained by Databricks. The endpoint https://
The Databricks MCP server requires appropriate API credentials and permissions based on the operations you want to perform. Check the documentation for specific permission requirements.
Yes, you can configure the same MCP server across Cursor, Claude Desktop, Cline, and other MCP-compatible clients simultaneously. Each client maintains its own configuration.
The MCP server itself is free to use, but you may need a Databricks account and API access, which may have its own pricing tiers. Check https://www.databricks.com for current pricing information.
The Databricks MCP server enables AI agents to manage tasks, projects, and workflows. Connect external MCP clients (Claude, Cursor, Replit) to Databricks MCP servers to query data, access Unity Catalog tables and vector indexes. OAuth or PAT; workspace-specific URL. This allows you to interact with Databricks's platform through natural language commands in your AI assistant.
Ensure your API key is correctly configured and your network allows outbound connections to databricks.com. Verify that the MCP endpoint URL https://<workspace-hostname>/api/2.0/mcp is accessible from your network.
Verify your API key is valid and has not expired. Generate a new key from https://www.databricks.com if needed. Make sure you're using the correct authentication method as specified in the documentation.
Check Databricks's status page for any ongoing incidents. Try restarting your MCP client and verify that the server endpoint is correct. If issues persist, consult the official documentation or support channels.
Explore other official MCP servers that work well with Databricks:
The Model Context Protocol (MCP) is an open standard that enables AI models to securely connect to external APIs, tools, and data sources. This official MCP server from Databricks provides standardized access to their services, allowing AI agents to interact with Databricks's platform seamlessly.
To manage all your MCP servers in one place, use Apigene's MCP Gateway, which provides centralized management, security, and governance for all your MCP server connections.
Learn more about MCP in our complete guide or browse our FAQ for common questions.
Domain
databricks.com
Website
https://www.databricks.comMCP Endpoint
https://<workspace-hostname>/api/2.0/mcp