CodeDNA

Problem Every developer has opened an unfamiliar codebase and asked the same questions: Why does this function exist? Who wrote it and why? What breaks if I change it? The answers are buried across hundreds of commits, scattered comments, and the memories of people who may no longer be on the team. There is no tool that surfaces this institutional knowledge quickly — so developers waste hours reading git logs, tracing call stacks, and guessing at risk before making changes. Solution CodeDNA gives any GitHub repository a living memory. Paste a repo URL and within seconds the system clones it, parses every file and function, maps the full commit history, and stores everything in a structured database. From that point forward, developers can interrogate the codebase the way they would interrogate a colleague who has been there from day one. The platform has four core capabilities. The Timeline view transforms raw commits into a clean yearly narrative — automatically categorised as features, bugfixes, security patches, optimisations, or milestones, with the business area inferred from the files touched. The File Explorer lets developers browse every file and function with extracted JSDoc descriptions, so they can orient themselves in minutes instead of hours. The Why Analysis uses Gemini AI to read a function's source code alongside its file's commit history and produce a grounded explanation of why it was introduced — complete with real commit evidence and a confidence rating. The Impact Analysis runs a static scan of the entire repo to find every caller of a function, then asks Gemini to assess the risk across Security, API, Database, Performance, Testing, and Architecture dimensions before a single line is changed. Tying it all together is Ask CodeDNA — a multi-turn AI chat agent that answers natural language questions about the repository. It is grounded in real data: the actual file tree, real commit messages, and the full conversation history for the session. It does not guess. If it lacks information, it says so. Target Users CodeDNA is built for developers joining an existing project, engineers preparing to refactor legacy code, and technical leads doing pre-change risk reviews. It is equally useful for open source contributors trying to understand a new codebase before submitting a pull request. What Makes It Unique Most AI coding tools generate new code. CodeDNA does the opposite — it reads existing code and explains it. The dependency graph is built through real static analysis, not AI inference, so there is no hallucination in the structural data. The AI layer is strictly used for narrative and risk assessment, always anchored to verified evidence. This hybrid approach — deterministic analysis feeding grounded AI — is what separates CodeDNA from a simple chatbot wrapper over a codebase. How It Works in Practice A developer opens CodeDNA, pastes a GitHub URL, and within moments has a full interactive map of the repository. They click a suspicious function, read why it was introduced three years ago during a security incident, see every file that calls it, and get an AI risk assessment before touching it. What used to take a day now takes two minutes.