Features
Everything you need to go from idea to infrastructure.
AI-Driven Design
Describe your infrastructure in plain English. OpenArchFlow's multi-provider AI engine — supporting Google Gemini 2.0/2.5, AWS Bedrock, and local WebLLM — instantly generates interactive architecture diagrams. Refine with follow-up chat: 'Add a CloudFront distribution in front of that S3 bucket.'
- Google Gemini 2.0 Flash, 2.5 Flash & Pro
- AWS Bedrock support
- Local WebLLM (runs in your browser)
- Incremental diagram modification via chat
- AI-generated documentation
AI-Driven Design
Infrastructure as Code
Export any diagram to Terraform (main.tf, variables.tf and outputs.tf). Dependencies are inferred from your diagram's edges across 50+ AWS resource types, and an optional AI-enhanced mode uses the HashiCorp Terraform MCP server for schema-accurate HCL. Treat the output as a solid scaffold and review it before you apply.
- 50+ AWS resource mappings
- Dependencies inferred from diagram edges
- main.tf, variables.tf and outputs.tf
- Monaco HCL editor with syntax highlighting
- AI-enhanced generation via the HashiCorp Terraform MCP server

Local Deployment with MiniStack
Launch a local AWS-compatible environment in about two seconds with MiniStack and deploy your diagram with one click. Each resource gets an interactive console: browse S3 objects, invoke Lambda functions, scan DynamoDB tables, read SQS messages, publish to SNS and EventBridge, test API Gateway routes and tail CloudWatch Logs.
- Deploy All with live per-node status badges
- 14 deployable service types, from S3 and Lambda to Glue
- Simulation traffic routed to the real deployed resources
- Configurable Traffic Source node (req/s)
- One-click teardown — zero cloud costs, no AWS account

AWS Glue on MiniStack
Run AWS Glue ETL jobs natively: MiniStack executes them in a local container built from the official AWS Glue 4.0 libraries. Glue Studio inside OpenArchFlow gives you a catalog browser, a table explorer, a PySpark editor, job runs with live Spark logs and an Athena query tab.
- Native job execution through MiniStack and Docker
- Catalog and Tables explorer
- PySpark editor and job runs
- Live JVM and Spark logs via local CloudWatch Logs
- Athena SQL over real S3 data using DuckDB (:full image)

Real-Time Collaboration
Work on diagrams with your team in real time over peer-to-peer WebRTC. Sessions are end-to-end encrypted: the key lives in the URL fragment and is never sent to the signaling server. Optionally back up diagrams to Google Drive with a one-click Google sign-in.
- Peer-to-peer real-time sync (Yjs + WebRTC)
- End-to-end encryption with zero-knowledge signaling
- Live presence in the toolbar
- Optional Google Drive auto-backup (drive.file scope)
Real-Time Collaboration
AWS Resource Discovery
Connect to your AWS account and import live resources from 24 services. OpenArchFlow infers connections from ARN references and renders them as an editable diagram you can export right away.
- 24 supported AWS services
- Connections inferred from ARN references
- AWS SSO login or access keys
- Selective import; per-service "No access" tolerance
- Export the discovered architecture to Terraform
AWS Resource Discovery
Cost Estimation
See cost estimates as you design. Powered by the AWS Price List API, they update when you add or remove resources — before you write a single line of Terraform.
- Live data from the AWS Price List API
- Configurable region, instance type and usage
- Bill of Materials panel for the whole architecture
- Export the estimate to CSV
Cost Estimation
Component Library
AWS, Azure, cloud-native, observability and integration components out of the box, with auto-layout, AWS architecture groups (VPC, subnet, AZ, region), zoom and keyboard shortcuts.
- Hundreds of AWS, Azure, cloud-native, observability and integration components
- Pre-built templates for AWS, Azure and generic architectures
- Auto-layout (dagre) and AWS architecture groups
- Export to PNG, SVG, PDF, SQL DDL and JSON
- Architecture simulation with a laser-pointer presentation mode

How OpenArchFlow Compares
| Feature | OpenArchFlow | draw.io | Lucidchart | Cloudcraft |
|---|---|---|---|---|
| AI Generation from Text | Partial | |||
| Terraform HCL Export | Partial | |||
| Local Deployment | ||||
| P2P Collaboration (E2EE) | Partial | |||
| AWS Resource Discovery | Partial | Partial | ||
| Real-Time Cost Estimation | Partial | |||
| Privacy-First (Local Storage) | Partial | |||
| Open Source | ||||
| Free | Partial |
Based on public information as of November 2026. "Partial" means limited, or available only on paid plans (for example, live AWS import in Lucid's Cloud Accelerator or Cloudcraft Pro). Check each vendor's current plans before deciding.