Docs · Quickstart

Five-minute quickstart.

Sign up, get a workspace, point a client at the gateway, watch the first policy evaluation light up the dashboard. No SDK to install. No code change. One config flip.

1. Sign up

Create an account at app.relayone.ai/sign-up. Free tier is fine; you can upgrade later. You'll be issued a default workspace.

2. Get your gateway URL and API key

From the dashboard's Setup tab, copy:

  • Gateway base URL: https://gateway.relayone.ai/v1/<workspace-slug>
  • API key: ro_live_... (treat like an OpenAI key; rotate via the dashboard)

3. Point a client at the gateway

Replace the baseURL on your existing OpenAI/Anthropic/Bedrock client with the gateway URL. Append the API key as Authorization: Bearer ro_live_.... Send your existing AI request. Done.

Example with the OpenAI Python SDK:

from openai import OpenAI

client = OpenAI(
    base_url="https://gateway.relayone.ai/v1/your-workspace",
    api_key="ro_live_xxxxxxxxxxxxxxxx",
)

resp = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Hello, RelayOne."}],
)
print(resp.choices[0].message.content)

The gateway forwards the request to OpenAI, evaluates policy inline, signs the receipt, and returns the response on the same TCP connection. Nothing about your application code changes.

4. Watch the dashboard

Open the Inspector tab. Your request appears in the live feed within seconds, with policy evaluation breakdown, cost attribution, and a link to the signed receipt. Filter by user, model, or policy outcome.

5. Add a policy

Open Policies, click New, and pick a starter (PII redaction, model allowlist, spend cap, geographic reroute). Save. The next request you send is evaluated against it. The dashboard tags the policy decision on the live feed.

6. What's next