What is a code graph?
A code graph is a structured representation of a codebase where functions, types, and modules are nodes, and relationships — calls, imports, dependencies, effects — are edges. Instead of reading files one at a time, a tool or AI agent queries the graph: who calls this, what does it touch, what breaks if it changes.
Why AI agents need one
Coding agents burn most of their context window reading files to rebuild an understanding that existed the last time someone — or something — analyzed the repository. A code graph makes that understanding persistent and queryable, so the agent asks a targeted question and receives a compact, structured answer instead of raw source.
The practical effect is fewer tokens spent orienting and more spent working, with answers — callers, dependencies, tests, contracts — that are computed rather than guessed.
Blast radius: the killer query
The question every engineer asks before a risky change — what does this touch? — becomes a computable answer: the transitive set of code affected through calls, contracts, shared state, and history. Severity-ranked impact turns 'I think it's safe' into a checked claim.