> ## Documentation Index
> Fetch the complete documentation index at: https://www.aptible.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# LLM Gateway

> Access 100+ models across Anthropic, OpenAI, and open-weight providers through a single, compliant API.

![LLM Gateway Overview](https://mintcdn.com/aptible/-KyDjlHfA6fZD2Z3/images/gateway-page-w-bg.png?fit=max\&auto=format\&n=-KyDjlHfA6fZD2Z3\&q=85\&s=09fe1263b00f678b0475b8634d84fc0e)

## Overview

> [!NOTE]
>
> The LLM Gateway works as a standalone product or alongside your existing Aptible infrastructure. No apps or databases required. Sign up and start sending requests right away. See the [pricing page](https://www.aptible.com/pricing) for plan and usage details.

> [!WARNING]
>
> **v1.0 keys reach end of life on December 7, 2026**, with brownouts on November 23 and November 30, 2026. See [Upgrading from v1.0](https://www.aptible.com/docs/llm-gateway/upgrading-from-v1.md) to migrate your keys.

Aptible's LLM Gateway gives you a single, compliant API for leading model providers, so your team can build securely with AI without having to configure directly with LLM providers. Through one endpoint and one key, you can:

* Access 100+ models across Anthropic, OpenAI, and open-weight providers through a single API
* Integrate with any OpenAI- or Anthropic-compatible SDK, and every major agent harness and coding assistant
* Govern which models your team can use with organization-wide model access policies
* Control and monitor spend with an organization-wide spend limit, plus per-key cost visibility
* Review request history and inspect full request and response logs to troubleshoot and refine prompts
* Send PHI safely under Aptible's HIPAA BAA, with automatic audit logging and encryption

## Get Started

> [card] [Claude Code](https://www.aptible.com/docs/getting-started/secure-llm-usage/claude-code.md)
>
> Route Claude Code through the LLM Gateway.

> [card] [OpenAI-compatible SDKs](https://www.aptible.com/docs/getting-started/secure-llm-usage/openai-sdk.md)
>
> Point any OpenAI-compatible SDK at the LLM Gateway.

> [card] [Anthropic-compatible SDKs](https://www.aptible.com/docs/getting-started/secure-llm-usage/anthropic-sdk.md)
>
> Point any Anthropic-compatible SDK at the LLM Gateway.

## Supported Models

The LLM Gateway supports 100+ models across Claude, GPT, and open-weight model families. [See Supported Models for full details](https://www.aptible.com/docs/llm-gateway/supported-models.md).

> [card] [Claude](https://www.aptible.com/docs/llm-gateway/supported-models.md)
>
> Anthropic's Haiku, Sonnet, and Opus models

> [card] [GPT & o-series](https://www.aptible.com/docs/llm-gateway/supported-models.md)
>
> OpenAI's GPT-4o, GPT-4.1, GPT-5, o3, o4, and more

> [card] [Open-weight models](https://www.aptible.com/docs/llm-gateway/supported-models.md)
>
> Amazon Nova, Meta Llama, Qwen, DeepSeek, and more

## Integrations & Compatibility

Call the LLM Gateway through OpenAI- or Anthropic-compatible endpoints, or point a supported agent harness or coding assistant at your own dedicated base URL and never edit it again.

[Learn more about Supported APIs →](https://www.aptible.com/docs/llm-gateway/supported-apis.md)

## Features

### Security & Compliance

The LLM Gateway is designed so compliance is built in at the infrastructure layer, not something your team has to configure or maintain. You route requests through the gateway, the controls apply automatically.

| Principle | How it works |
| - | - |
| **Centrally managed credentials** | Your raw provider API keys never leave Aptible. Developers and applications authenticate with scoped LLM keys. Rotate or revoke a key without touching any application code or provider settings. |
| **Model access controls** | Your organization has an allowlist of approved models. Any model not on the list is blocked for every key in the organization. Unapproved models are structurally unavailable, not just discouraged. |
| **Full audit trail** | Every request and response is logged automatically — prompt content, response, model, key, token counts, cost, and timestamp. No configuration required. |
| **No training of models on PHI** | Providers accessed through the gateway are contractually prohibited from using your data for model training. This is enforced at the infrastructure layer, not by relying on provider-side policy. |
| **HIPAA BAA coverage** | Aptible's BAA covers all models and providers accessed through the gateway. You don't need separate BAAs per provider, and you don't need a dedicated stack to send PHI. |
| **Spend controls** | A hard monthly spending cap prevents runaway costs from a misconfigured agent or a compromised key. |
| **LLM High Availability** | Aptible automatically fails over to another available provider if a model's primary provider has an outage. |

[Learn more about Security & Compliance →](https://www.aptible.com/docs/llm-gateway/security-and-compliance-guardails.md)

### Model Access Policies

Control which models your team can use with a single, organization-wide model access policy. The policy applies to every LLM key in your organization, giving you centralized control over model usage across your applications and developers.

[Learn more about Model Access Policies →](https://www.aptible.com/docs/llm-gateway/model-access-policies.md)

### Cost Visibility & Control

![Edit Monthly Spend Limit](https://mintcdn.com/aptible/p1TivY-miYPAedOd/images/edit-monthly-spend-limit.png?fit=max\&auto=format\&n=p1TivY-miYPAedOd\&q=85\&s=d3013319b46f05dd633d53426ca8499d)

Set an organization-wide monthly spending limit, plus track LLM usage at the organization and per-key level. So you always know what's being spent and where.

[Learn more about Cost Visibility & Control →](https://www.aptible.com/docs/llm-gateway/cost-control.md)

### Audit Logging

![Logs Ui](https://mintcdn.com/aptible/q-D0a8Ga9TUzSuGk/images/logs-ui.png?fit=max\&auto=format\&n=q-D0a8Ga9TUzSuGk\&q=85\&s=5ee8e67ed1283fb37698611276177f86)

Every LLM request and response is automatically logged. View request history and full request and response logs in the Aptible dashboard, with every log retained by Aptible for 6+ years.

[Learn more about Audit Logging →](https://www.aptible.com/docs/llm-gateway/audit-logging.md)

## Support and Feedback

We'd love to hear from you as you use the LLM Gateway. If you have questions, run into issues, or have feature requests, [contact us](https://app.aptible.com/support).