---
title: AI Runtime Control
summary: null
url: https://www.fastly.com/documentation/reference/api/ai-runtime-control
---

AI Runtime Control (ARC) is a gateway for managing and monitoring traffic to AI model providers. These APIs allow you to create and manage virtual keys, configure provider connections, and inspect usage metrics and session logs for AI requests made through the ARC gateway.

- [Gateway](https://www.fastly.com/documentation/reference/api/ai-runtime-control/gateway/) - The <a href="https://www.fastly.com/documentation/reference/api/ai-runtime-control">AI Runtime Control</a> gateway forwards your requests to AI model providers. You authenticate to the gateway with a <a href="https://www.fastly.com/documentation/reference/api/ai-runtime-control/virtual-keys">virtual key</a>, and the gateway uses the credentials configured on that key to reach the provider for you. Most gateway endpoints accept the same requests and return the same responses as the provider APIs OpenAI Chat Completions, OpenAI Responses APIs, Anthropic Messages API, and Amazon Bedrock Runtime <code>InvokeModel</code> and <code>InvokeModelWithResponseStream</code> APIs. This means you can keep using your existing provider client and change only its base URL and credentials. Specify models by name, for example <code>gpt-4</code> or <code>claude-sonnet-4-20250514</code>. Bedrock endpoints work differently. They put the Bedrock model ID or ARN in the request path instead of a <code>model</code> field, and authenticate with a <code>Bearer</code> token instead of signing requests with AWS Signature Version 4 (SigV4). Each endpoint follows its provider's request and response format, so see that provider's documentation for the full list of supported fields.
- [Provider Connections](https://www.fastly.com/documentation/reference/api/ai-runtime-control/provider-connections/) - These endpoints allow you to create and manage provider connections, which supply the credentials <a href="https://www.fastly.com/documentation/reference/api/ai-runtime-control">AI Runtime Control</a> uses to reach an AI provider on your behalf. A provider connection is authenticated either with a provider API key or, for Amazon Bedrock, an AWS Identity and Access Management (IAM) role. Each customer account is limited to 25 provider connections.
- [Providers](https://www.fastly.com/documentation/reference/api/ai-runtime-control/providers/) - These endpoints return reference data about the AI providers and models <a href="https://www.fastly.com/documentation/reference/api/ai-runtime-control">AI Runtime Control</a> supports, and setup information for connecting to Amazon Bedrock.
- [Sessions](https://www.fastly.com/documentation/reference/api/ai-runtime-control/sessions/) - This endpoint returns session logs for requests made through <a href="https://www.fastly.com/documentation/reference/api/ai-runtime-control">AI Runtime Control</a>, including request and response data and metadata, for debugging and audit purposes.
- [Usage Metrics](https://www.fastly.com/documentation/reference/api/ai-runtime-control/usage-metrics/) - These endpoints return usage metrics for requests made through <a href="https://www.fastly.com/documentation/reference/api/ai-runtime-control">AI Runtime Control</a>, including request counts, session counts, and token consumption. Metrics can be filtered by virtual key, provider, and model.
- [Virtual Keys](https://www.fastly.com/documentation/reference/api/ai-runtime-control/virtual-keys/) - These endpoints allow you to create and manage virtual keys, which are used to authenticate requests made through the <a href="https://www.fastly.com/documentation/reference/api/ai-runtime-control">AI Runtime Control</a> gateway. Each virtual key can be configured with request-per-minute and token-per-minute rate limits, a spend-limit flag, and a failover policy.