Set up

​

Before you start

• A Guardway account (sign up) with the owner or admin role
• An API key from at least one LLM provider (OpenAI, Anthropic, Google, Groq, …)
• Outbound HTTPS (443) to gateway.guardway.ai from wherever your applications run. If a model endpoint or MCP server you connect only accepts listed IP addresses, allow the SaaS Gateway’s address on it. See Network Requirements.

​

The five steps

1

Open the SaaS tab

In the dashboard, go to Gateways and select the SaaS tab (marked Beta).

2

Click Setup SaaS Gateway

One click — no tokens, no configuration choices. The card flips to Provisioning while Guardway creates your dedicated gateway. It typically shows Online in one to two minutes.

3

Connect a provider

Open Configuration → Providers, pick a preset, paste your provider API key, and attach it to your SaaS gateway. Your gateway picks the change up automatically within about a minute.Full walkthrough →

4

Create a Guardway API key

Open Configuration → API Keys and create a key. This is the key your applications send — your provider keys stay server-side and are never exposed to callers.Full walkthrough →

5

Make your first request

The gateway is OpenAI-compatible: point any OpenAI SDK at the endpoint and use your Guardway API key.

curl

Python

Node

curl https://gateway.guardway.ai/v1/chat/completions \
  -H "Authorization: Bearer YOUR_GUARDWAY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o-mini",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Use a model your connected provider serves (for Google, e.g. gemini-2.5-flash; for OpenAI, e.g. gpt-4o-mini).

​

Try it without writing code

​

If something looks off

• Stuck on Provisioning past ~5 minutes — click into the card later or contact support; provisioning reports its status automatically.
• {"error":"no providers available"} — no provider is attached to this gateway yet, or it was attached moments ago; give it a minute (step 3).
• 401 from the endpoint — the API key is wrong, revoked, or from a different organization.
• Provider errors (e.g. 429 quota) — passed through verbatim from your LLM provider; check that account’s plan and limits.