patrickchugh/terravision

What it solves

TerraVision automates the creation of professional cloud architecture diagrams for AWS, Azure, and Google Cloud. It eliminates the need to manually draw diagrams in tools like Lucidchart or Visio, ensuring that visual documentation stays perfectly synced with the actual infrastructure code (Terraform) or can be generated instantly from a plain-English description via an AI assistant.

How it works

TerraVision operates as a local tool that can be integrated into AI assistants (like Claude, ChatGPT, and Cursor) via the Model Context Protocol (MCP) or as a plugin. It works in two primary directions:

1. Design to Code: An AI assistant takes a natural language description, generates a JSON graph, and TerraVision renders it into a diagram. The assistant can then write the corresponding Terraform code based on that design.
2. Code to Diagram: TerraVision analyzes terraform plan output to generate an accurate visual representation of the resources, VPCs, subnets, and zones being deployed.

It uses Graphviz for rendering and can export diagrams to PNG, SVG, editable draw.io files, or interactive HTML pages.

Who it’s for

Cloud architects, DevOps engineers, and developers who use Terraform to manage infrastructure and want to maintain accurate, professional architecture diagrams without manual effort.

Highlights

• AI-Native Integration: Works as an extension/plugin for Claude Desktop, ChatGPT, and other MCP-compatible clients.
• Provably Accurate: Generates diagrams directly from Terraform code, making the code the single source of truth.
• Llocal & Secure: Runs entirely on the user's machine with no cloud access required and no code uploads.
• Multi-Cloud Support: Official icons and grouping conventions for AWS, Azure, and GCP.
• CI/CD Automation: Includes a GitHub Action to automatically redraw diagrams whenever Terraform code changes.
• Interactive Output: Supports clickable HTML diagrams for security audits and documentation.