# Create AI requests and add them to your collections

You create AI requests to interact with an AI model of your choice. Choose any OpenAI-compatible model, whether it's from [OpenAI](https://platform.openai.com/docs/), [Anthropic](https://docs.anthropic.com/), [Google](https://ai.google.dev/gemini-api/docs), or an alternative provider such as [HuggingFace](https://huggingface.co). Use AI requests to experiment, test, and evaluate different AI models, whether they're fine-tuned, self-hosted, or local. Use the model that best fits your security requirements, budget, and workflow.

Then, add a [Model Context Protocol](https://modelcontextprotocol.io/introduction) (MCP) server to your request to give the model added context. For example, add a server that performs an action or gets information from a trusted data source.

## Create your AI requests

Create AI requests to interact with AI models. Then, add them to your collections and organize them in your workspace for your project.

To learn more, see [Create an AI request for your collection](/docs/postman-ai/ai-requests/create/).

## Add MCP servers

Add MCP servers to your AI requests to give the AI models you interact with further context.

To learn more, see [Add an MCP server to your AI request](/docs/postman-ai/ai-requests/add-mcp-servers/).

## Interact with AI models

Use your AI requests to interact with AI models. As you interact with the model, the model returns a response. You can use these responses to experiment, test, and evaluate an AI model.

To learn more, see [Use your AI request to interact with an AI model](/docs/postman-ai/ai-requests/interact/).

## Manage your AI request settings

Edit your authentication details and adjust how you interact with an AI model.

To learn more, see [Manage how you interact with an AI model](/docs/postman-ai/ai-requests/manage/).