Gateway AI

One endpoint for every frontier model

Gateway AI is an LLM gateway that gives your team access to OpenAI, Anthropic, and Google Gemini through a single API. Every request is authenticated, rate-limited, cost-tracked, and audit-logged — with zero prompt retention.

You get below-market pricing on every model because we aggregate demand and pass bulk discounts directly to you.

Why Gateway AI?

  • Save money. Pay below standard provider rates on every API call. See exactly how much you saved in your dashboard.
  • Stay compliant. Append-only, immutable audit log on every request. Zero prompt data retained. Access policies enforced at the gateway.
  • One integration. Connect to OpenAI, Anthropic, and Gemini through a single gateway. Manage all your API keys, budgets, and usage in one place.

Two Ways to Integrate

Native Mode

Point your existing provider SDK at Gateway AI. Change two lines of config — your code stays exactly the same.

1from openai import OpenAI
2
3client = OpenAI(
4 api_key="sk-proxy-...", # your Gateway AI key
5 base_url="https://api.trygateway.ai/openai/v1" # your Gateway AI gateway
6)
7
8# everything else is identical
9response = client.chat.completions.create(
10 model="gpt-4o",
11 messages=[{"role": "user", "content": "Hello"}]
12)

Native mode supports every feature of each provider’s API: streaming, function calling, vision, tool use, and more. Nothing is translated or modified.

Unified Mode

Use the OpenAI SDK to call any provider. Switch models by changing a string.

1from openai import OpenAI
2
3client = OpenAI(
4 api_key="sk-proxy-...",
5 base_url="https://api.trygateway.ai/v1"
6)
7
8# Call Anthropic with the OpenAI SDK
9response = client.chat.completions.create(
10 model="anthropic/claude-sonnet-4-6",
11 messages=[{"role": "user", "content": "Hello"}]
12)

Unified mode translates requests and responses between OpenAI’s format and each provider’s native format, so you can use one SDK for everything.

Architecture

Your Code (any SDK)
Gateway AI Gateway
├── Authenticate (proxy key)
├── Enforce policies (rate limits, budgets, IP allowlists)
├── Route to provider
├── Track usage & cost
└── Log audit trail
LLM Provider (OpenAI / Anthropic / Gemini)

Next Steps

Ready to start? The Quick Start guide gets you from zero to your first request in under 60 seconds.