Enola indexes multi-repo codebases into a knowledge graph for AI agents
Enola is an open-source engine that indexes a codebase — even one split across multiple repositories like an iOS app, Android app, backend, and frontend — into a single persistent . It runs as an MCP server, letting it directly answer questions like "what calls this function," "what breaks if I change this interface," and "is this code actually used." The project grew out of two developers' experience building a golf app whose codebase spread across several repos, a common pain point once a project splits into .
The core idea is that doesn't change between sessions, yet currently re-derive it from scratch every time using grep and guesswork. Pre-building the as a graph means agents don't have to rediscover it repeatedly, and multiple repositories can be merged into one combined graph.
Key points
- Indexes code spread across multiple repos (iOS, Android, backend, frontend) into one persistent
- Runs as an MCP server so can query it directly
- Answers questions like "what calls this," "impact of an interface change," "is this code reachable"
- Open source
- Aims to stop AI agents from re-discovering from scratch each session