Don’t want to run infrastructure? The SaaS Gateway (Beta) gives you a fully managed gateway in about two minutes — no Docker, no deployment. The steps below are for the self-hosted gateway.
Before you start
- A machine or VM with Docker installed (see Requirements)
- An API key from at least one LLM provider (OpenAI, Anthropic, Groq, etc.)
- Outbound HTTPS (port 443) from that machine to Guardway and your provider’s API. See Network Requirements for the hosts to allow.
The six steps
1
Sign up for Guardway
Create your account at app.guardway.ai. You’ll receive an invite email, set a password, and land on the dashboard.Full walkthrough →
2
Deploy a gateway
In the dashboard, open Gateways and click Register New Gateway to generate a one-time registration token. Paste it into the
docker run command the dialog gives you and start the container on any host in your network.Full walkthrough →3
Activate the gateway
Within a minute the gateway sends a heartbeat and appears as Online in the Gateways list. Click Connect to point the dashboard at its local URL.Full walkthrough →
4
Connect a provider
Open Configuration → Providers, pick a preset (OpenAI, Anthropic, Groq, etc.), paste your API key, and attach it to your gateway.Full walkthrough →
5
Sync models
Open Configuration → Models and click Sync Models on the provider. The gateway discovers what’s available and enables them for inference.Full walkthrough →
6
Test in the Playground, then check logs
Open the Playground, pick a model, send a message. Then open Monitoring → Logs to see the request and Monitoring → Traces to see how it flowed through guardrails, routing, and the provider.Playground → · Logs & Traces →