Databricks MCP Server

1# Databricks MCP Server
2 
3A Model Context Protocol (MCP) server that connects to Databricks API, allowing LLMs to run SQL queries, list jobs, and get job status.
4 
5## Features
6 
7- Run SQL queries on Databricks SQL warehouses
8- List all Databricks jobs 
9- Get status of specific Databricks jobs
10- Get detailed information about Databricks jobs
11 
12## Prerequisites
13 
14- Python 3.7+
15- Databricks workspace with:
16  - Personal access token
17  - SQL warehouse endpoint
18  - Permissions to run queries and access jobs
19 
20## Setup
21 
221. Clone this repository
232. Create and activate a virtual environment (recommended):
24   ```
25   python -m venv .venv
26   source .venv/bin/activate  # On Windows: .venv\Scripts\activate
27   ```
283. Install dependencies:
29   ```
30   pip install -r requirements.txt
31   ```
324. Create a `.env` file in the root directory with the following variables:
33   ```
34   DATABRICKS_HOST=your-databricks-instance.cloud.databricks.com
35   DATABRICKS_TOKEN=your-personal-access-token
36   DATABRICKS_HTTP_PATH=/sql/1.0/warehouses/your-warehouse-id
37   ```
385. Test your connection (optional but recommended):
39   ```
40   python test_connection.py
41   ```
42 
43### Obtaining Databricks Credentials
44 
451. **Host**: Your Databricks instance URL (e.g., `your-instance.cloud.databricks.com`)
462. **Token**: Create a personal access token in Databricks:
47   - Go to User Settings (click your username in the top right)
48   - Select "Developer" tab
49   - Click "Manage" under "Access tokens"
50   - Generate a new token, and save it immediately
513. **HTTP Path**: For your SQL warehouse:
52   - Go to SQL Warehouses in Databricks
53   - Select your warehouse
54   - Find the connection details and copy the HTTP Path
55 
56## Running the Server
57 
58Start the MCP server:
59```
60python main.py
61```
62 
63You can test the MCP server using the inspector by running 
64 
65```
66npx @modelcontextprotocol/inspector python3 main.py
67```
68 
69## Available MCP Tools
70 
71The following MCP tools are available:
72 
731. **run_sql_query(sql: str)** - Execute SQL queries on your Databricks SQL warehouse
742. **list_jobs()** - List all Databricks jobs in your workspace
753. **get_job_status(job_id: int)** - Get the status of a specific Databricks job by ID
764. **get_job_details(job_id: int)** - Get detailed information about a specific Databricks job
77 
78## Example Usage with LLMs
79 
80When used with LLMs that support the MCP protocol, this server enables natural language interaction with your Databricks environment:
81 
82- "Show me all tables in the database"
83- "Run a query to count records in the customer table"
84- "List all my Databricks jobs"
85- "Check the status of job #123"
86- "Show me details about job #456"
87 
88## Troubleshooting
89 
90### Connection Issues
91 
92- Ensure your Databricks host is correct and doesn't include `https://` prefix
93- Check that your SQL warehouse is running and accessible
94- Verify your personal access token has the necessary permissions
95- Run the included test script: `python test_connection.py`
96 
97## Security Considerations
98 
99- Your Databricks personal access token provides direct access to your workspace
100- Secure your `.env` file and never commit it to version control
101- Consider using Databricks token with appropriate permission scopes only
102- Run this server in a secure environment
103