Docs

Getting started

Quickstart

Deploy the RenLayer proxy, register an agent, and route your first OpenAI call through governance, in roughly five minutes.

This quickstart takes a fresh tenant from zero to a fully governed agent call in about five minutes. You will deploy the proxy, register an agent, generate an API key, and route an OpenAI request through the governance layer.

Prerequisites

  • A RenLayer tenant (request access from the console).
  • Access to a host that can run the proxy (Docker, Kubernetes, or a Linux VM).
  • An upstream API key for the model provider you want to govern (e.g. OpenAI).

1. Pull and run the proxy

The proxy ships as a single container. Set the database URL and the platform API endpoint, then start it.

docker run -d --name renlayer-proxy \
  -e RENLAYER_DATABASE_URL="postgres://..." \
  -e RENLAYER_API_URL="https://api.renlayer.com" \
  -e RENLAYER_LISTEN_ADDR="0.0.0.0:8080" \
  -p 8080:8080 \
  ghcr.io/renlayer/proxy:latest

The proxy is now listening on port 8080 and is ready to receive agent traffic.

2. Register an agent

In the console, open Agents and create a new agent. Give it a name (e.g. support-bot-v1), pick the upstream provider (OpenAI), and save. The console returns an agent ID and an API key scoped to that agent.

Treat the API key like a service credential: store it in your secret manager, never in source.

3. Point your agent at the proxy

Replace the upstream base URL in your agent code with the proxy URL, and use the RenLayer API key as the bearer token.

from openai import OpenAI

client = OpenAI(
    base_url="https://proxy.your-org.internal/v1",
    api_key="rl_live_...",  # RenLayer agent API key
)

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

The OpenAI SDK is unmodified, only the base_url and api_key change.

4. Verify in the console

Open the Sessions page in the console. Within seconds you should see a new trace with the agent name, the model, latency, token counts, and the request body. This is the audit trail that policies, DLP, and the audit log all build on.

5. Add your first policy

In Policies, create a rule that flags any prompt containing the keyword salary. Re-run the agent with a prompt that mentions salaries, then refresh Sessions: the new trace will be marked FLAGGED and surfaced in the dashboard.

Where to go next

Last updated: