marmotdata/marmot

What it solves

Marmot addresses the difficulty of discovering and understanding data assets across a team's infrastructure. It eliminates the complexity of enterprise data catalogs by providing a lightweight, single-binary solution for cataloging tables, topics, queues, and APIs, ensuring that both human teams and AI agents have a clear, certified context of the data landscape.

How it works

It functions as a metadata-first context layer. Users can catalog various data assets via a plugin system and enrich them with business context, ownership, and shared glossaries. The system provides full-text search, structured queries, and interactive lineage to trace data flows. To integrate with AI, it uses the Model Context Protocol (MCP) and an API to expose this certified metadata to AI agents, preventing them from guessing about the data structure.

Who it’s for

Data engineers, architects, and teams who need a streamlined data discovery tool that is easy to deploy and integrates directly with their AI agents.

Highlights

• AI-Ready Context: Uses MCP to provide AI agents with real-time, certified metadata.
• Single Binary Deployment: Simplifies setup by avoiding complex infrastructure requirements.
• Interactive Lineage: Allows users to trace data flows from source to destination for impact analysis.
• Broad Asset Support: Catalogs everything from databases and message queues to APIs and dashboards.