Dremio Model-Context Protocol (MCP)

This project implements a Model-Context Protocol (MCP) server for Dremio. It provides a standardized way to interact with Dremio databases through the MCP.

Features

Resources

Exposes Dremio data through Resources:

• dremio://catalogs - Lists all catalogs
• dremio://schemas/{catalog} - Lists all schemas in a catalog
• dremio://tables/{catalog}/{schema} - Lists all tables in a schema
• dremio://schema/{table_name} - Gets schema information for a table

Tools

• list_catalogs

  • Lists all available catalogs in Dremio
  • Usage: list_catalogs()

• list_schemas

  • Lists all schemas in a specific catalog
  • Usage: list_schemas("catalog_name")

• list_tables

  • Lists all tables in a specific schema
  • Usage: list_tables("catalog_name", "schema_name")

• execute_sql

  • Executes a SQL query against Dremio
  • Returns results in a structured format with metadata
  • Usage: execute_sql("SELECT * FROM catalog.schema.table LIMIT 10")

Requirements

• Python 3.10+
• Dremio instance (Cloud or Software)
• PyArrow with Flight support

Installation

Using pip

pip install -e .

Using uv (recommended)

# Install uv if you don't have it
curl -sSf https://astral.sh/uv/install.sh | bash

# Install the package
uv pip install -e .

Configuration

Create a config.yaml file in the root directory with the following configuration:

# Dremio connection settings
dremio:
  # Connection type: 'software' for Dremio Software or 'cloud' for Dremio Cloud
  type: software

  # Dremio Software connection details
  software:
    hostname: localhost
    port: 32010
    username: your_username
    password: your_password
    use_ssl: false

# Query settings
query:
  # Maximum number of rows to return in a query result
  max_rows: 10000
  # Maximum number of columns to return in schema request
  max_columns: 1000
  # Timeout in seconds for query execution
  timeout: 300

Environment Variables

You can also configure the server using environment variables:

DREMIO_HOSTNAME=localhost
DREMIO_PORT=32010
DREMIO_USERNAME=your_username
DREMIO_PASSWORD=your_password
DREMIO_USE_SSL=false

For Dremio Cloud:

DREMIO_TYPE=cloud
DREMIO_ENDPOINT=data.dremio.cloud
DREMIO_PAT_TOKEN=your_personal_access_token

Usage

Running the Server

# If installed from package
mcp-dremio-server

# Or run directly
python run_mcp_server.py

Integration with Claude or Other MCP Clients

Add the following configuration to your MCP client settings:

MacOS

// ~/Library/Application\ Support/Claude/claude_desktop_config.json
{
  "mcpServers": {
    "dremio_service": {
      "command": "/Users/USERNAME/.local/bin/uv",
      "args": [
                "--directory",
                "/Users/USERNAME/PycharmProjects/dremio-python-mcp",
                "run",
                "run_mcp_server.py"
            ]
    }
  }
}

Windows

// %APPDATA%/Claude/claude_desktop_config.json
{
  "mcpServers": {
    "dremio-mcp": {
      "command": "mcp-dremio-server",
      "args": []
    }
  }
}

Example Usage

Listing Catalogs and Schemas

// List all catalogs
list_catalogs()

// List schemas in a catalog
list_schemas("my_catalog")

// List tables in a schema
list_tables("my_catalog", "my_schema")

Executing SQL Queries

// Execute a query
results = execute_sql("SELECT * FROM my_catalog.my_schema.my_table LIMIT 10")

Security Features

• SQL Validation: Prevents execution of non-read-only queries
• Input Validation: Prevents SQL injection via table/schema names
• TLS Support: Secure communication with Dremio
• Authentication: Support for username/password and token authentication

Debugging

Since MCP servers run over stdio, debugging can be challenging. For the best debugging experience, use the MCP Inspector:

npx @modelcontextprotocol/inspector mcp-dremio-server

Upon launching, the Inspector will display a URL that you can access in your browser to begin debugging.

License

This project is licensed under the MIT License - see the LICENSE file for details.