AxiomCode

AxiomCode is the team publishing AxiomCode Graph through the AxiomCodeAI GitHub organization. Its public work centers on helping engineers and coding agents understand the structure of software before making changes. The team’s website connects that code-context work with a broader interest in formal verification.

AxiomCode Graph maps relationships in a codebase so callers, dependencies, and affected tests can be explored explicitly. The project’s documentation explains its analysis approach and distinguishes language support and maturity, making the implementation more useful to inspect than a general promise of better coding.

A developer could use the graph to investigate an unfamiliar module, plan a cross-file refactor, or review which tests a change may reach. The published demonstrations show context queries and paths through real code. The deeper AxiomEngine offering is a separate layer rather than another name for the Graph repository.

Radar follows AxiomCode for making code understanding a concrete part of agent workflows. The useful evidence includes source, examples, and evaluation methods. Reported benchmark results remain specific to their tested languages and workloads; unresolved calls and incomplete coverage can leave relevant relationships outside a selected path.