Configure Azure API Management as an AI Gateway for AI models, MCP tools, and agents. WHEN: semantic caching, token limit, content safety, load balancing, AI model governance, MCP rate limiting, jailbreak detection, add Azure OpenAI backend, add AI Foundry model, test AI gateway, LLM policies, configure AI backend, token metrics, AI cost control, convert API to MCP, import OpenAPI to gateway.
Detected risks:
The azure-aigateway skill enables users to configure Azure API Management (APIM) as a centralized AI Gateway for managing AI models, MCP tools, and agents. It addresses challenges in controlling AI access, enforcing policies, and monitoring usage across multiple AI backends. By leveraging this skill, organizations can standardize AI interactions, apply governance rules consistently, and ensure operational compliance, all while integrating seamlessly with Azure services. This approach simplifies managing token limits, content safety, and load balancing, which are critical for scaling AI solutions securely and efficiently.
Key features include semantic caching for cost reduction, token rate limiting to manage expenses, content safety policies for compliance, and load balancing for optimal AI model utilization. Users can add AI backends like Azure OpenAI or AI Foundry models, apply comprehensive governance policies, and monitor token usage with built-in metrics. The skill also supports converting APIs to MCP tools, importing OpenAPI definitions into the gateway, and testing AI endpoints using Azure CLI commands. These capabilities provide observability, cost control, and streamlined administration of AI services through a single gateway.
This skill is ideal for organizations deploying multiple AI models and tools in Azure environments that require robust governance, monitoring, and compliance controls. Typical users include cloud architects, AI engineers, and platform administrators responsible for managing AI workloads at scale. It is particularly valuable for teams needing to enforce token usage limits, implement semantic caching, maintain content safety standards, and integrate AI models efficiently without building custom infrastructure from scratch.
Deployment of the AI gateway requires first preparing Azure API Management (APIM) using the azure-prepare skill. After APIM is deployed, the azure-aigateway skill can configure backends, apply policies, and enable monitoring.
This skill requires the Azure CLI (az) for configuration, backend management, and testing. Users must have appropriate permissions for APIM and any AI backend resources.
Yes, you can add multiple AI backends, including Azure OpenAI and AI Foundry models, and configure load balancing, authentication, and governance policies for each backend.
Cost management is achieved through token limits and semantic caching policies, which help control usage and reduce unnecessary calls to AI backends by 60-80%.
Testing requires valid subscription keys and properly configured APIM backends. Endpoints must follow the defined API paths and include necessary authentication headers.
Quick Setup:
.claude/skills/Repository
microsoft/github-copilot-for-azure