Cloud Weaver is an AWS architecture assistant that converts a plain-English idea into a validated architecture design.

It supports a full architecture workflow:

• prompt-to-diagram generation (Mermaid)
• rule-based validation (syntax + architecture constraints)
• session-aware context for multi-turn refinement
• optional CloudFormation generation after approval

• Frontend (CloudFront): https://d2k45vy1qt3ioe.cloudfront.net/
  • Website has been shutdown due to Hackathon ending
• Backend (API Gateway WebSocket): wss://9vihcpxj86.execute-api.us-west-2.amazonaws.com/dev/

• What this project does
• End-to-end flow
• System architecture
• AWS services users can design with
• AWS services used to build this application
• Validation pipeline
• Session memory model
• Repository layout
• Getting started
• Configuration
• Current project focus

Cloud Weaver helps teams quickly go from idea to deployable architecture artifacts while keeping output constrained to an approved AWS service set.

Primary outputs:

• Mermaid architecture diagram text (rendered in frontend)
• architecture reasoning/explanations
• CloudFormation YAML draft generated from approved architecture context

1. A user submits a system idea in the frontend chat UI.
2. The frontend sends the request over API Gateway WebSocket routes.
3. Lambda orchestrates prompt construction and model invocation.
4. Amazon Bedrock returns architecture reasoning + Mermaid graph.
5. Backend validators enforce Mermaid syntax and architecture constraints.
6. Validated output is returned to the frontend for rendering.
7. If approved, CloudFormation generation is requested using the latest session context.
8. CloudFormation output is syntax-validated before deployment-oriented workflows continue.

• React + Vite chat application
• WebSocket client for low-latency request/response flow
• Markdown + Mermaid rendering for assistant output

• Python Lambda handlers for orchestration
• Bedrock integration for generation
• DynamoDB-backed session/message persistence
• Validation modules for Mermaid and CloudFormation outputs

• Mermaid syntax validation
• Architecture rule validation
• CloudFormation YAML structure/syntax validation

7 Predefined selectable services for generated architectures:

• Amazon Bedrock
• AWS Lambda
• Amazon S3
• API Gateway
• CloudFront
• DynamoDB
• AWS IAM

These services power Cloud Weaver itself:

Validation is separated to reduce invalid model output before deployment actions:

• validation/mermaid_syntax.py
  • header and structural checks
  • quote/delimiter balance checks
• validation/architecture_rules.py
  • approved service enforcement
  • connection integrity and duplicate/orphan checks
• validation/cloudformation_syntax.py
  • YAML syntax checks
  • root object/resource structure checks

Cloud Weaver uses sessionID for multi-turn session continuity.

• Frontend stores sessionID in browser localStorage.
• Backend loads recent session history from DynamoDB and injects it into prompt context.
• CloudFormation generation references the latest approved architecture context from the same session.

Current Lambda environment values:

• CHAT_HISTORY_LIMIT=20
• MESSAGES_TABLE=Messages
• SESSION_TABLE=Sessions
• SESSION_TTL_SECONDS=604800

Note: backend supports both SESSION_TABLE and SESSIONS_TABLE variable names.

aws-architect/
├── backend/
│   ├── lambda/
│   │   ├── handler.py
│   │   ├── bedrock_client.py
│   │   ├── session_store.py
│   │   └── test_local.py
│   ├── validation/
│   │   ├── architecture_rules.py
│   │   ├── mermaid_syntax.py
│   │   └── cloudformation_syntax.py
│   └── requirements.txt
├── frontend/
│   ├── src/
│   │   ├── App.jsx
│   │   ├── MermaidChart.jsx
│   │   └── components/
│   └── package.json
└── prompts/

• Node.js 20+
• Python 3.12
• AWS credentials configured locally (for backend Bedrock/AWS calls)

cd frontend
npm install
npm run dev

Optional clean reinstall after dependency drift:

cd frontend
rm -rf node_modules
npm ci

cd backend
python3.12 -m venv .venv
source .venv/bin/activate
python -m pip install --upgrade pip
python -m pip install -r requirements.txt

Optional local harness:

cd backend/lambda
python3 test_local.py

Frontend runtime configuration (typically in .env under frontend/):

• VITE_WS_URL for WebSocket backend endpoint
• VITE_TEST_MODE=true for local fallback behavior

Backend configuration relies on Lambda environment variables and AWS IAM permissions for Bedrock, DynamoDB, and API integrations.