Skip to content

Deployment

Deploy LLM Council to the cloud in under 5 minutes with one-click deployment, or run locally with Docker Compose.

Quick Deploy

Platform Deploy Best For
Railway Deploy on Railway Production use, webhook integrations
Render Deploy to Render Evaluation, free tier testing

Render Cold Start

Render Free tier spins down after 15 minutes of inactivity. This causes 30-60 second cold starts that may timeout n8n webhooks. For reliable webhook integration, use Railway or Render paid tier.

Platform Comparison

Feature Railway Render Free Render Paid
Cold Start None 30-60s after 15min idle None
Free Tier $5/mo trial 750 hrs/mo $7/mo+
Webhook Ready Yes No (timeout risk) Yes
Auto-scaling Yes No Yes

Deployment Options

  • :material-train: Railway

    Recommended for production. No cold starts, template marketplace discovery, and webhook reliability.

  • :material-palette: Render

    Free tier for evaluation. Great for testing before committing to a paid plan.

  • :material-docker: Local Docker

    Run locally with Docker Compose. Perfect for development and testing.

Required Configuration

All deployment options require these environment variables:

Variable Required Description
OPENROUTER_API_KEY Yes Your OpenRouter API key for LLM calls
LLM_COUNCIL_API_TOKEN Yes Bearer token for API authentication
PORT No Server port (default: 8000)

Security Notice

Never commit API keys to your repository. Use platform secrets management:

  • Railway: Project Settings > Variables
  • Render: Service > Environment
  • Local: .env file (gitignored)

Verify Deployment

After deployment, verify your instance is running:

# Health check (no auth required)
curl https://your-instance.railway.app/health

# Test API (requires auth)
curl -X POST https://your-instance.railway.app/v1/council/run \
  -H "Authorization: Bearer YOUR_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"prompt": "What is 2+2?"}'

Next Steps