Instalar
Instala la CLI localmente. Codna puede correr en tu máquina, tu nube o infraestructura auto-hospedada.
pip install codna codna --version
Docs
Instala Codna con un solo comando.
CLI primero, MCP listo y compatible con la GitHub App nativa. Empieza con la comprensión del repo local, luego conecta correcciones y PRs.
Inicio rápido
Primero, comprende un repo. Luego pide a Codna que corrija un issue específico.
codna triage . codna fix . --issue "the checkout test is failing"
CLI
Los comandos principales son intencionalmente pequeños y scriptables.
codna triage . --json codna fix . --issue "..." --tests codna fix https://github.com/org/repo --issue "..."
Servidor MCP
Añade Codna a Cursor o Claude en una línea. Tu agente puede pedir a Codna mapas de símbolos, dependencias, impacto en tests y paquetes de evidencia enfocados.
codna mcp start --repo .
GitHub App nativa
Instala la GitHub App para que Codna haga triage de issues, comente con evidencia de causa raíz y abra pull requests de corrección.
# App flow Install codna → select repos → enable fix PRs
Configuración
Bring your own model key, or use Codna's managed model — it's the same Codna either way.
model: provider: openai key: env:CODNA_MODEL_KEY privacy: egress: fail-closed redact_secrets: true
FAQ
¿Codna entrena con mi código? No. Codna está diseñado para no entrenar con código de clientes.
¿Puedo auto-hospedarlo? Sí. El auto-hospedaje está disponible para siempre, con opciones on-prem y air-gapped con soporte para Enterprise.
Frequently asked
Install with a single command from your terminal. The codna cli runs locally, so you bring your own API key and your code never leaves your machine unless you choose otherwise.
The MCP server for coding integrates natively with Cursor and Claude. Add it to your editor's MCP config and Codna's graph engine becomes available to your AI assistant without any extra steps.
The GitHub App monitors your repository and opens pull requests with verified fixes automatically. Every fix is validated by your own test suite before the PR is created.
No. Codna is self-hostable, and egress is fail-closed by default. You supply your own API key, and the system is designed so your code is never used for model training.
The deterministic graph engine maps the relevant code first, producing an evidence bundle measured at around 600 tokens. Head-to-head against Cursor, Codna used 5× fewer tokens and ran 1.7× faster, every fix test-verified (87/87).
Codna supports 250+ languages, and has mapped 130 repositories in 9.2 seconds for zero LLM tokens. If your project compiles or resolves dependencies, Codna can graph it.