Deploy to Render

Render provides a fully managed platform with auto-TLS, private networking, and managed PostgreSQL. The Ferro Labs AI Gateway ships with a render.yaml Blueprint that provisions everything in one click.

One-click deploy

The Blueprint provisions:

• Web service — Docker-based gateway with health checks
• PostgreSQL database — Managed Postgres for API key storage, config history, and request logging
• Auto-wired DSNs — All three store backends (API keys, config, request logs) connect to the database automatically
• Generated MASTER_KEY — A secure master key is generated during setup

You will be prompted for your OPENAI_API_KEY (and any other provider keys you need).

What the Blueprint configures

The render.yaml in the gateway repository sets up:

Resource	Type	Details
ferrogw	Web Service (Docker)	Exposes port 8080, health check at /health
ferrogw-db	PostgreSQL	Managed database for all persistence

Environment variables are auto-wired:

Variable	Source	Purpose
MASTER_KEY	Auto-generated	Admin authentication
OPENAI_API_KEY	User-provided	OpenAI provider key
API_KEY_STORE_BACKEND	postgres	API key storage backend
API_KEY_STORE_DSN	Database internal URL	Connection string
CONFIG_STORE_BACKEND	postgres	Config storage backend
CONFIG_STORE_DSN	Database internal URL	Connection string
REQUEST_LOG_STORE_BACKEND	postgres	Request log storage backend
REQUEST_LOG_STORE_DSN	Database internal URL	Connection string

Manual setup

If you prefer to configure Render manually instead of using the Blueprint:

Step 1: Create a PostgreSQL database

In the Render dashboard, create a new PostgreSQL database. Note the Internal Connection String — you will need it for the environment variables.

Step 2: Create a web service

1. Click New → Web Service
2. Connect your GitHub repository or use the public image: ghcr.io/ferro-labs/ai-gateway:latest
3. Set the environment to Docker
4. Configure environment variables:

MASTER_KEY=fgw_your-master-key
OPENAI_API_KEY=sk-your-key
PORT=8080

# Storage backends — all point to the Postgres database
API_KEY_STORE_BACKEND=postgres
API_KEY_STORE_DSN=postgresql://user:pass@host:5432/dbname
CONFIG_STORE_BACKEND=postgres
CONFIG_STORE_DSN=postgresql://user:pass@host:5432/dbname
REQUEST_LOG_STORE_BACKEND=postgres
REQUEST_LOG_STORE_DSN=postgresql://user:pass@host:5432/dbname

warning

Never hardcode API keys in source files. Always use Render's environment variable management, which encrypts values at rest.

Step 3: Configure health check

Set the health check path to /health so Render can monitor the service and restart it if it becomes unhealthy.

Step 4: Deploy

Click Create Web Service. Render builds the Docker image and starts the gateway. You will see the service URL in the dashboard once the deploy completes.

Adding more providers

Add provider API keys as environment variables in the Render dashboard:

ANTHROPIC_API_KEY=sk-ant-...
GEMINI_API_KEY=...
MISTRAL_API_KEY=...

No code changes or redeployment needed — the gateway picks up new keys on restart. Use the Manual Deploy button or update an environment variable to trigger a fresh deploy.

Verify the deployment

# Health check
curl https://your-service.onrender.com/health

# Test request
curl https://your-service.onrender.com/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $MASTER_KEY" \
  -d '{
    "model": "gpt-4o",
    "messages": [{"role": "user", "content": "Hello from Render!"}]
  }'

Cost considerations

Resource	Render Plan	Approximate Cost
Web service	Starter	$7/month
PostgreSQL	Starter	$7/month
Total		~$14/month

Render's free tier is available for testing but has limitations (spins down after inactivity, limited compute). For production workloads, use the Starter plan or above.

Related pages

• Deploy to Railway — Alternative one-click cloud deployment
• Docker Compose — Self-hosted deployment with PostgreSQL
• Kubernetes — Helm chart for production clusters
• Server settings — All environment variables reference