OpenEZ Graph is a local-first code intelligence tool that helps index the entire codebase and documentation into a knowledge graph that can be searched and reused, supporting AI coding agents to retrieve the correct context instead of having to read the entire source each time they work. The project is designed to be lightweight, running locally on SQLite, without dependence on cloud, Postgres, or Redis, and can be operated flexibly through CLI, MCP server, or web dashboard.

OpenEZ Graph currently supports full-text search, graph expansion, code context, memory to retain technical decisions, workspace management, incremental indexing, and watch mode, significantly saving tokens, time, and effort when working with large codebases.

Additionally, the tool has an MCP server to connect with AI coding tools such as Codex, Claude Code, OpenCode, allowing agents to find the correct file, symbol, dependency relationships, and relevant contexts without needing to scan the entire project again.

The web UI is also available for users to check the workspace, explore the graph, and visually view local indexed data. OpenEZ Graph is still under development, and we look forward to receiving feedback, testing, and contributions from the community to continue improving the project.