SentinAI Setup Guide
Quick Start (Local Development)
npm install
# Copy environment template
cp .env.optimism.example .env.local # Optimism / OP Stack
npm run dev
Visit http://localhost:3002 (or port shown in terminal).
Environment Variables
Minimum Required (3 variables for full functionality)
L2_RPC_URL=https://your-l2-rpc-endpoint.com # L2 Chain RPC
ANTHROPIC_API_KEY=your-api-key-here # AI features
AWS_CLUSTER_NAME=my-cluster-name # K8s (auto-detects K8S_API_URL & region)
Note:
K8S_API_URLandAWS_REGIONare auto-detected at runtime fromAWS_CLUSTER_NAME.
AWS credentials use the standard chain: env vars,~/.aws/credentials, or IAM Role.
Template Example
The repository includes a .env template:
.env.optimism.example— Optimism / OP Stack
Copy the template to .env.local and customize.
EC2 Deployment (Automated)
For production deployment on AWS EC2:
bash scripts/install.sh
This script handles:
- Dependency installation
- Environment setup
- Service configuration
- Firewall rules
See EC2 Setup Guide for detailed instructions.
Next Steps
After setup:
- Run a demo: Demo Scenarios
- Configure scaling: Autonomy Cockpit User Guide
- Production operations: Daily Operations Runbook
Simulation Mode
By default, SentinAI runs in simulation mode (dry-run). Scaling decisions are logged but not executed on the cluster.
To enable live execution:
- Set
SCALING_SIMULATION_MODE=falsein.env.local - Ensure AWS credentials have EKS write permissions
- Review Autonomy Cockpit User Guide for safety controls
Troubleshooting
Connection Issues
- Verify
L2_RPC_URLis accessible - Check AWS credentials:
aws sts get-caller-identity - Confirm EKS cluster name matches
AWS_CLUSTER_NAME
AI Features Not Working
- Verify
ANTHROPIC_API_KEYis valid - Check API key permissions (Claude Haiku 4.5 required)
- Review logs for rate limit errors
Scaling Not Triggered
- Confirm
SCALING_SIMULATION_MODE=falsefor live execution - Check cooldown period (5 minutes by default)
- Review Testing Guide
For more issues, see Redis Setup and component-specific guides.