How to Self-Host LLM Gateway

Want full control over your LLM infrastructure? Self-host LLM Gateway on your own servers—keep your data in-house, avoid platform fees, and customize everything. Here's how to get running in under 5 minutes.

Option 1: Unified Docker Image (Easiest)

1docker run -d \
2  --name llmgateway \
3  --restart unless-stopped \
4  -p 3002:3002 -p 3003:3003 -p 3005:3005 -p 3006:3006 -p 4001:4001 -p 4002:4002 \
5  -v ~/llmgateway_data:/var/lib/postgresql/data \
6  -e AUTH_SECRET=your-secret-key-here \
7  ghcr.io/theopenco/llmgateway-unified:latest

1docker run -d \
2  --name llmgateway \
3  --restart unless-stopped \
4  -p 3002:3002 -p 3003:3003 -p 3005:3005 -p 3006:3006 -p 4001:4001 -p 4002:4002 \
5  -v ~/llmgateway_data:/var/lib/postgresql/data \
6  -e AUTH_SECRET=your-secret-key-here \
7  ghcr.io/theopenco/llmgateway-unified:latest

One command. All services. Running in seconds.

Tip: Pin to a specific release tag (e.g., v1.2.3) in production to avoid unexpected updates.

Option 2: Split Services via Docker Compose

For more control over individual services (useful for scaling or debugging):

1git clone https://github.com/theopenco/llmgateway.git
2cd llmgateway
3cp .env.example .env
4# edit .env
5docker compose -f infra/docker-compose.split.yml up -d

1git clone https://github.com/theopenco/llmgateway.git
2cd llmgateway
3cp .env.example .env
4# edit .env
5docker compose -f infra/docker-compose.split.yml up -d

Access Your Instance

Once running, your services are available at:

What You Get

Self-hosting gives you:

• Zero platform fees — No percentage taken from your API spend
• Data sovereignty — All requests stay on your infrastructure
• Unlimited customization — Modify the codebase to fit your needs
• Same features — Analytics, caching, and routing work just like the hosted version

For the full setup guide with environment configuration and production tips, see the Self-Host documentation.