Skip to main content

Install AI Gateway

The Levo AI Gateway is a proxy that sits between your applications and LLM providers (OpenAI, Anthropic, etc.). It gives you full observability into AI traffic and lets you enforce guardrails — blocking prompt injection, PII leakage, toxic content, and more — without any changes to your application code.

Routes, backends, and guardrail policies are configured in the Levo dashboard and automatically applied to the gateway.

1. Prerequisites​

Before installing the Levo AI Gateway, ensure you have:

  • Network connectivity from your host to the Levo SaaS API

Without guardrails:

  • CPU: At least 2 cores
  • RAM: At least 1 GB

With guardrails (ML models):

  • CPU: At least 4 cores
  • RAM: At least 8 GB
  • Storage: At least 15 GB (~6.3 GB of ML models, plus images and headroom)

2. Retrieve Your Authorization Key​

Refer to Generate a CLI Authorization Key for instructions.

3. Retrieve Your Environment ID​

Your Environment ID is not yet available in the Levo dashboard. Please contact support@levo.ai to obtain it.

4. Follow instructions for your platform​

5. Configure Guardrails​

Installing the gateway gives you observability. Enforcement comes from the policies you configure on top of it.

AI Guardrails covers what the scanners detect, how a policy is expressed, and how to operate them — including the one configuration mistake that makes a fully configured deployment scan nothing.

Was this page helpful?