> This page location: Backend > AI Gateway > Get started > Overview
> Full Neon documentation index: https://neon.com/docs/llms.txt

> Summary: Neon AI Gateway is the LLM inference layer built into the Neon backend. One Neon credential gives you access to models across multiple providers. Standard AI SDKs work without code changes. Each branch gets its own gateway endpoint.

# Neon AI Gateway

One API for all frontier & open-source models. At-cost rates.

Neon AI Gateway is the LLM inference layer built into the Neon backend. It lets you call models from Anthropic, OpenAI, Google, and other providers using your Neon credential, without setting up separate provider accounts. Your existing OpenAI or Anthropic SDK works without code changes. Just point it at your branch endpoint.

During the private preview, AI Gateway is available for **new projects** in the **AWS us-east-2** region only, and foundation model access requires a paid Neon plan.

**Important:** Participation in this Private Preview is subject to our Private Preview Terms. Access is not available to users, organizations, or entities located in or operating from regions restricted by Anthropic's [Supported Regions Policy](https://www.anthropic.com/supported-countries). This restriction also applies to entities that are majority owned, directly or indirectly, by companies headquartered in unsupported regions.

- **One credential for all providers.** A single Neon credential gives you access to models from Anthropic, OpenAI, Google, Meta, DeepSeek, Databricks, and Alibaba. No separate provider accounts needed.
- **Standard SDKs, one URL change.** OpenAI SDK, Anthropic SDK, and google-genai all work out of the box.
- **AI follows your branches.** Each branch has its own gateway endpoint. If you use Neon branches for preview deployments, AI requests from a feature branch are scoped to that branch. It's the same isolation your database already gets.
- **Streaming support.** Server-sent events work on all endpoints with no extra configuration.

## Quickstart

- [Quickstart](https://neon.com/docs/ai-gateway/get-started): Get a credential and make your first inference request in minutes.
- [Models](https://neon.com/docs/ai-gateway/models): Browse the full model catalog and learn how to specify models in requests.
- [Chat completions](https://neon.com/docs/ai-gateway/chat-completions): Use the OpenAI-compatible endpoint with any model in the catalog.
- [Authentication](https://neon.com/docs/ai-gateway/authentication): Understand how Neon credentials work with AI Gateway.

## Starter templates

Browse working examples at [build-on-neon.vercel.app](https://build-on-neon.vercel.app/). Two templates use AI Gateway:

**`ai-sdk`**: An image-generation agent that routes model calls through AI Gateway, stores results in Neon Storage, and writes metadata to Postgres on a Neon Function.

```bash
neonctl bootstrap --template ai-sdk
```

**`mastra`**: A personal assistant that uses AI Gateway for LLM calls with Postgres-backed memory on a Neon Function.

```bash
neonctl bootstrap --template mastra
```

---

## Related docs (Get started)

- [Quickstart](https://neon.com/docs/ai-gateway/get-started)
- [Models](https://neon.com/docs/ai-gateway/models)