What is OpenArchFlow? OpenArchFlow is an open-source initiative that harnesses the power of Artificial Intelligence (specifically Google's Gemini models) to convert natural language descriptions directly into AWS architecture diagrams. The premise is simple yet powerful: you describe your requirements—“I need a Go application running on ECS with an RDS database and an ElastiCache cluster”—and the tool automatically generates the corresponding visual representation. The Technological Edge: Precision via MCP One of the greatest challenges in using AI for architecture is ensuring it utilizes the latest components and patterns. To solve this, OpenArchFlow employs a modern context-aware architecture: Native AI: The project utilizes Google's Gemini models for high-level semantic interpretation. Edge AI with WebLLM and Phi-3 (Microsoft): The project isn't limited to cloud-based models. We have implemented support for WebLLM to run the Phi-3 model locally, directly in the browser. This enables a hybrid experience: the efficiency of local processing combined with the intelligence of an optimized LLM. AWS MCP Servers: Unlike other tools, we integrate MCP (Model Context Protocol) servers that expose official AWS documentation directly to the model. This ensures the AI has access to precise, up-to-date service definitions during diagram generation. Diagrams as Code: The output is processed to generate coherent, standardized technical visualizations.

OpenArchFlow

AI-Powered AWS Architecture Diagram Generator

AI-Powered AWS Architecture Diagram Generator