Core Highlights of OpenArchFlow

OpenArchFlow is an open-source progressive web application (PWA) that quickly generates interactive AWS architecture diagrams through natural language descriptions. Key features include:

• Zero-configuration & privacy-first: No registration required, data processed locally
• Dual AI modes: Cloud (Google Gemini 2.5 Flash) and local (WebLLM + Phi-3) options available
• Real-time cost estimation: Integrated with AWS Pricing API
• Rich interactive features: Auto-layout, architecture simulation, multi-format export
• Open-source & extensible: MIT license supports community contributions

Project Background & Core Positioning

Created by developer Rafael Sales, targeting cloud architects, DevOps engineers, and developers.

Core Positioning

• Zero setup: No registration needed, use directly in the browser
• Privacy-first: Data processed locally, no sensitive information uploaded
• Tech stack: Next.js 16+, TypeScript, Tailwind CSS, React Flow (diagrams), Zustand (state management)

Core Features & Technical Implementation

Natural Language to Architecture Diagram

Input a description (e.g., "Serverless API with Lambda, API Gateway, DynamoDB, S3"), and the AI generates the diagram in seconds. Dual modes:

• Cloud AI: Google Gemini 2.5 Flash (fast)
• Local AI: WebLLM running Phi-3 (offline & private)

Real-time Cost Estimation

Integrated with AWS Pricing API; customize region/instance/usage to get accurate cost estimates

Component Library & Interaction

• Hundreds of components: Covering AWS, Azure, cloud-native, etc.
• Features: Drag-and-drop layout, dagre auto-optimization, architecture simulation, laser pointer presentation, export to PNG/JSON, multi-diagram management

Incremental Generation & Conversational Design

Incremental Generation

AI intelligently modifies/extends existing architectures, supporting iterative optimization

Diagram Chat

Discuss with AI assistant:

• Get design explanations
• Pricing estimates
• Security review suggestions Lower the barrier for non-professional users to participate

Technical Architecture & Extensibility

Architecture

PWA architecture supports offline use; local AI runs via WebGPU

Open-source Contribution

• MIT license (auditable code)
• Contribution method: Fork → feature branch → PR; follow Conventional Commits standards, changes recorded in CHANGELOG.md

Application Scenarios & Value

Applicable Scenarios

• Rapid prototyping: Generate initial architecture before review meetings
• Technical documentation: Generate professional diagrams
• Cost evaluation: Compare costs of different solutions
• Team collaboration: Standardize diagram-based communication
• Learning AWS: Understand service combinations via natural language

Value

Improve design efficiency, reduce documentation costs; open-source for customization and extension

Future Development Directions

Plans:

• Add more AWS services (AppSync, EKS, ECS, etc.)
• Optimize UI/UX
• Support internationalization
• Improve mobile experience
• Enhance unit/integration testing