apache/doris-mcp-server

What it solves

It allows AI agents and MCP (Model Context Protocol) hosts to interact with Apache Doris databases in a secure, read-only manner. It bridges the gap between LLMs and complex database capabilities by providing a structured way for agents to discover and execute database operations without granting them full administrative access.

How it works

The server implements the MCP protocol to expose Apache Doris capabilities as a set of tools. It uses a hierarchical discovery system where agents first call a "domain" (like doris_query or doris_cluster) to discover specific child capabilities and their schemas. It supports multiple transport methods, including stdio and Streamable HTTP, and enforces a strict security boundary where Doris RBAC (Role-Based Access Control) remains the final authority on data visibility.

Who it’s for

Developers building AI agents or using MCP-compatible hosts that need to query, monitor, or analyze data stored in Apache Doris.

Highlights

• Hierarchical Tooling: Organizes 55 child capabilities into 8 stable domains (e.g., catalog, query, cluster, pipeline, search, governance, lakehouse, and semantic).
• Read-Only Security: Built-in read-only catalog to prevent accidental or malicious data modification.
• Enterprise-Grade Security: Supports JWT, OAuth/OIDC, and Doris-backed OAuth with redacted secrets and errors.
• Detailed Observability: Provides capabilities for cluster metrics, slow query diagnosis, and lineage tracking.
• Flexible Deployment: Supports both local stdio and network-based HTTP transports.