MCP Directory
Back to Blog
aihawkmcp-clientfeder-crbrowser-agent

AIHawk MCP Client: AI Browser Agent for Claude Code & Gemini CLI

October 21, 2026
2 min read

AIHawk MCP Client: AI Browser Agent for Claude Code & Gemini CLI

AIHawk is an open-source Python-based MCP Client designed to act as an AI browser agent, enabling web automation through plain English commands. It provides a real browser environment for AI assistants to interact with the web, performing actions like browsing, clicking, typing, and reading content.

Integrating AIHawk with Your Assistant via MCP

The core utility of AIHawk lies in its MCP integration, allowing various AI assistants to leverage its web browsing capabilities. The setup involves installing AIHawk and then registering it as a tool with your chosen assistant.

First, install the necessary dependencies. For Windows using PowerShell:

powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
$env:Path = "$env:USERPROFILE\.local\bin;$env:Path"
uvx invisible-playwright fetch

On Linux, use the following bash commands:

curl -LsSf https://astral.sh/uv/install.sh | sh
source $HOME/.local/bin/env
uvx invisible-playwright fetch

Once installed, you can register AIHawk with your assistant. The --scope user stealth flag indicates that AIHawk should operate in a "stealth" mode, likely referring to its browser agent capabilities.

For Claude Code:

claude mcp add --scope user stealth -- uvx aihawk

For Codex:

codex mcp add stealth -- uvx aihawk

For Gemini CLI:

gemini mcp add --scope user stealth uvx aihawk

After these steps, your assistant will recognize and be able to invoke AIHawk to perform web-based tasks.

Capabilities as a Web Browsing Agent

AIHawk's primary function is to translate natural language instructions into concrete web actions. It operates a real browser, allowing it to:

  • Browse the web: Navigate to specified URLs.
  • Click elements: Interact with buttons, links, and other clickable components.
  • Type text: Fill out forms, search bars, and input fields.
  • Read web content: Extract information from web pages.

This enables developers to delegate complex web interactions to their AI assistants, using simple English prompts rather than writing explicit automation scripts.

References

Back to all posts