Repo understanding without model tokens.
Codna parses symbols, imports, call paths, tests, and dependencies into a live graph the agent can query.
Codna maps your repo in ~60ms for zero tokens. Then an agent ships a fix your own tests verify — roughly pennies each.
Codna parses symbols, imports, call paths, tests, and dependencies into a live graph the agent can query.
Instead of dumping files into a context window, Codna creates a compact bundle: suspect files, call chain, failing test, and risk map.
bundle: failing_test: checkout.spec.ts suspect_files: 4 call_paths: 7 estimated_context: ~600 tokens
Core capabilities
Map any local path or git URL in ~60ms for zero LLM tokens, and see where the change belongs.
Fix from a ~600-token evidence bundle — root cause, confidence, and regression risk attached.
Review generated changes with blast radius, tests touched, and API impact.
Open pull requests via the GitHub App — verified by your own test suite, with the evidence attached.
Distribution
A deterministic engine builds a dependency and blast-radius graph in about 60ms, using zero LLM tokens. That graph produces a focused ~600-token evidence bundle — 162x less context than reading the repository — so the AI agent works only on what matters.
Every fix is verified by your own test suite before it ships. Nothing merges until your tests pass.
Codna supports 250+ languages, and has mapped 130 repositories in 9.2 seconds for zero tokens. If your project has tests, Codna can work with it.
In head-to-head testing across 1,000+ tasks, Codna used 5× fewer tokens than Cursor and ran 1.7× faster, with every fix verified by the project's own tests (100%). Both agents were measured on the same tasks.
Codna ships as a CLI, an MCP server that works inside Cursor and Claude, and a native GitHub App that opens verified fix pull requests directly in your repo.
No. You can self-host Codna, bring your own API key, and egress is fail-closed. Your code is never used for training.