Data Olympus is an open-source knowledge base and Model Context Protocol (MCP) server harness created by knaisoma. Built in Python using FastMCP and SQLite, the tool is designed to manage context, documentation, and structured project knowledge for AI coding agents such as Claude Code, GitHub Copilot, and other MCP-compliant clients. In standard workflows, Data Olympus runs as a local MCP server that exposes tools and structured context to coding assistants. The underlying storage relies on SQLite with FTS5 full-text indexing for retrieval, with vector embeddings available as an optional opt-in feature. Agents query the harness to inspect health metrics (such as pending knowledge base items), fetch canonical project rules, and search project documentation directly from the terminal or IDE. What differentiates Data Olympus from generic agent harnesses is its adherence to standardized agent instruction structures (such as canonical AGENTS.md files with thin pointers for CLAUDE.md and GitHub Copilot). Instead of requiring complex external database infrastructure, it runs lightweight SQLite and FastMCP instances that integrate into Python development environments with minimal overhead. The project is available as open-source software under GitHub version control, actively maintained using modern Python tooling like uv.
Tags: ai agent, cli, developer tools, memory, OpenSource