Minimalist pipeline: Linux console + Python + local inference runtime + open-source model. From a structured prompt, the model produces a complete Terraform script including resources, dependencies, and configuration shells. The system is then orchestrated in CI/CD pipelines or shell scripts for deployment. Goal: reduce manual error and standardize infrastructure configurations while keeping the LLM out of the cloud.

Highlights

  • Open-source LLM in local inference (no external API)
  • Terraform main.tf generation from structured prompt
  • Inclusion of dependency configuration shells
  • Integrable into CI/CD pipelines or automation shells
  • Reduction of manual error in IaC writing
  • No project data leaves the local machine.

Demonstrative prototype. Reusable pattern for IaC automation contexts.