Apache Pinot

Apache Pinot

Python-based MCP server for Apache Pinot enabling metadata access and read-only queries.

14
Stars
6
Forks
4
Releases

Overview

Pinot MCP Server is a Python-based MCP server designed to interact with Apache Pinot. Built on the FastMCP framework, it exposes capabilities to retrieve Pinot metadata (tables, segments, and schemas), execute read-only SQL queries, and view index or column-level metadata from a Pinot cluster. The server is intended to support real-time analytics workflows and is designed to work with Claude Desktop, enabling business users to explore data, craft prompts, and perform metadata-driven analyses directly against Pinot. Configuration is done via a uvicorn-style .env file in the repository root (with an example file) and supports optional OAuth authentication for secure access. It can be run locally with uv --directory . run mcp_pinot/server.py and can be built and deployed as a Docker image. The project also supports DXT desktop extensions packaging to deliver tools to Claude Desktop. All MCP tooling is implemented in the Pinot class (utils/pinot_client.py), and the repository provides quick-start prompts to exercise the MCP server.

Details

Owner
startreedata
Language
Python
License
Apache License 2.0
Updated
2025-12-07

Features

Pinot metadata discovery

List tables, segments, and schema information from Pinot to quickly inspect cluster metadata.

Read-only SQL query execution

Execute read-only SQL queries against Pinot through the MCP interface.

Index and column-level metadata viewing

View index and column metadata for deeper data understanding.

Claude Desktop integration

Designed to work with Claude Desktop to enable real-time analytics and metadata queries for business users.

OAuth authentication (optional)

Optional OAuth-based authentication with configurable environment variables for secure access.

FastMCP-based server

Built on the FastMCP framework for a lightweight, extensible MCP server.

DXT desktop extensions support

Supports packaging and installing Claude Desktop extensions via DXT.

Docker and local development support

Includes instructions for building and running the container and local development workflow.

Audience

Business usersExplore Pinot datasets via Claude, run metadata and data queries in real time.
Data engineersIntegrate Pinot MCP server into Claude-driven workflows and developer tooling.
Analytics teamsEnable real-time analytics and metadata access against Pinot clusters for decision making.

Tags

PinotMCPClaude DesktopMetadataSQLOLAPFastMCPDXTOAuthDocker