# mcp-aoai-web-browsing

> MCP Server

This repository provides a server/client implementation using the Model Context Protocol (MCP) and Azure OpenAI, leveraging FastMCP and Playwright for testing. It bridges MCP responses to OpenAI function calling format

## Overview

- **Category:** AI
- **Language:** Python
- **Stars:** 34
- **Forks:** 1
- **Owner:** kimtth
- **GitHub:** https://github.com/kimtth/mcp-aoai-web-browsing
- **Created:** 2024-12-13T12:12:08+00:00
- **Updated:** 2025-03-28T01:16:50+00:00
- **Source:** https://model-context-protocol.com/servers/aoai-web-browsing-model-context-protocol

## Setup

## Setup

1.  Rename `.env.template` to `.env`, then fill in the values in `.env` for Azure OpenAI:

    ```bash
    AZURE_OPEN_AI_ENDPOINT=
    AZURE_OPEN_AI_API_KEY=
    AZURE_OPEN_AI_DEPLOYMENT_MODEL=
    AZURE_OPEN_AI_API_VERSION=
    ```

2.  Install `uv` for python library management

    ```bash
    pip install uv
    uv sync
    ```

3.  Execute `python chatgui.py`


## Tools

## Available Tools

1.  MCP Server (Built with `FastMCP`)
2.  `Playwright` (End to end testing framework)
3.  MCP response conversion to OpenAI function calling format
4.  `MCP-LLM Bridge` (Customizes the conversion of MCP server response to OpenAI function calling format)
5.  `playwright_navigate` (Navigate to a URL tool)