doris

Doris MCP Server is a Python-based backend that implements the MCP protocol to expose tooling for interacting with Apache Doris databases. It supports a suite of enterprise-grade capabilities, including secure token-based access, real-time configuration validation, hot-reloadable token management, session caching, and a web-based token management interface. Through the Tools Manager, you can discover and invoke a variety of tools for query execution, metadata extraction, analysis, and governance, all orchestrated via a unified MCP HTTP endpoint. The server is designed to work with Doris-backed data sources and can leverage prompt templates and analysis tools to support NL2SQL, query explain plans, and metadata discovery tasks. The HTTP interface supports both streaming and standard request/response modes for flexible integration with clients like Cursor and other MCP-enabled copilots. To start, install the package and run the provided CLI entrypoint, which will bind to your host and port and expose the set of tools registered by the server.

Typical workflows include: routing SQL execution through the query engine with caching and retry logic, using the prompts manager for data analysis prompts, and leveraging the resources manager to expose database metadata and catalog federation across Doris and external data sources. You can also administer tokens and bindings via the web-based management UI on localhost, enable multi-worker scaling for throughput, and rely on the integrated security framework to protect endpoints and SQL execution.

Prerequisites:

• Python 3.12 or newer
• pip (installed with Python)
• Access to a Doris database endpoint (host, port, user, password, database)

Installation steps:

1. Create a virtual environment (recommended)

python -m venv doris-env
source doris-env/bin/activate  # on Windows use doris-env\Scripts\activate

2. Install the Doris MCP Server package from PyPI

pip install doris-mcp-server

3. Verify the CLI entrypoint is available

doris-mcp-server --help

4. Start the server (example configuration)

doris-mcp-server \
  --transport http \
  --host 0.0.0.0 \
  --port 3000 \
  --db-host 127.0.0.1 \
  --db-port 9030 \
  --db-user root \
  --db-password your_password

5. Access the MCP API endpoint and the token management UI at http://localhost:3000 (or the configured host/port).

If you prefer running via a module-based invocation (instead of the CLI), you can also run the server with:

python -m doris_mcp_server

Notes and tips:

• Ensure the Doris database connection details are correct and that TLS/ACL settings align with your security policy.
• The v0.6.0 release introduces enterprise authentication with token-bound configurations; use the token management dashboard to create and revoke tokens, and bind tokens to specific databases or catalogs.
• Enable hot-reload for configuration changes without downtime by using the built-in configuration manager and token validation features.
• If you run into port conflicts or binding issues, verify firewall rules and ensure the chosen host/port are accessible to clients.
• The MCP server supports both streaming and standard HTTP modes; use streaming for high-throughput interactions and standard mode for simple request/response workflows.
• For production deployments, consider deploying behind a reverse proxy, enabling TLS, and configuring session pools and multi-worker settings to meet your load requirements.