Repowise is an open-source codebase intelligence tool for AI agents and humans: it indexes your code, call graph, git history, tests, and docs once, then answers questions about health, dead code, and change impact through MCP. It has earned more than 7,800 stars on GitHub and is under very active development, with updates in October 2026.
Repowise’s pitch is to understand your codebase without paying your agent to rediscover it. It indexes your code, call graph, git history, tests, documentation, and design decisions once, on your machine. Then you and your coding agent can ask it things: what calls this, what breaks if I change it, what is dead, what to fix first, why it was built this way, and what evidence stands behind each change that landed. On top of the index it computes code health scores, detects dead code, generates documentation, and produces git analytics and architectural decision records, all exposed to agents via the Model Context Protocol. According to the project, teams see measurably less redundant exploration when agents query the index instead of re-reading the repository on every task.
Repowise is free for one developer on a laptop, and the core is open source under the AGPL-3.0 license. It is built to run across a company’s repositories, and team or company-wide deployments are the commercial side of the model, so larger organizations should check the site for current terms.
Repowise is for engineering teams whose coding agents waste tokens re-learning the same repository, and for tech leads who need health, dead-code, and impact analysis grounded in evidence. Solo developers get a free local tool; larger teams get a shared codebase memory for their agents.