← Discover MCPs and Agents
g
MCPAI & MLGitHub

gai-cli

Links

Repository

README

From the repo.

gai-ag (Gemini Autonomous Agent)

gai-ag is a professional, fast, and intelligent command-line tool that brings the Google Gemini API to your terminal. It features an advanced "Agent" mode that can both answer questions directly and perform automatic modifications on your project.

  • Autonomous Agent Mode: Automatically fixes errors, runs tests, and generates solutions.
  • Project Brain: Creates a .gai/ folder in each project to store history, state, and errors.
  • Smart Scanning: Caches project structure and prioritizes critical files to save tokens.
  • Polyglot Support: Automatically detects Flutter, Node.js, and Python projects.

✨ Features

  • 🤖 Agent Mode: Analyzes files in your project, plans requested changes (code writing, file creation, deletion, moving), and applies them with your approval.
  • 💬 Interactive Chat: Communicate fluently with Gemini through a multi-modal chat interface.
  • 📁 Context Injection (@): Add files as context to your chat using @file.py or @src/.
  • 🎨 Premium UI: Stylish and readable output powered by the rich library.
  • 🌍 Multi-language Support: English and Turkish language options.
  • 🔒 Secure Operations: File system operations are restricted to the project directory.

🚀 Installation

From PyPI (Easiest)

pip install gai-ag

For Development

  1. Clone the repository:

    git clone https://github.com/bugraakdemir/gai-cli.git
cd gai-cli

  2. Install dependencies:

    pip install -e .

🛠️ Usage

Note: Both gai and gai-ag commands work in the terminal!

One-time Question

gai "What are list comprehensions in Python?"
# or
gai-ag "What are list comprehensions in Python?"

Interactive Mode (Chat & Agent)

Start the interactive mode by simply typing gai or gai-ag:

gai
# or
gai-ag

Collected info

  • ★ 0 stars
  • Language: Python
  • Source updated: 1/14/2026

Config for your environment

Replace {MCP_ENDPOINT_URL} with this MCP’s endpoint URL (from its repo or docs above). No API key — you connect directly.

Tool

OS

Config file: ~/.cursor/mcp.json

{
  "mcpServers": {
    "mcp-server": {
      "url": "{MCP_ENDPOINT_URL}"
    }
  }
}

Paste into mcpServers in the config file. Restart Cursor after saving.

If this MCP is also published on mcpchannel.ai, you can subscribe from Browse and use the gateway config there instead.