Skip to content

Home

Jump to bottom
D.Sanjai Kumar edited this page Jul 7, 2025 · 4 revisions

Welcome to the official wiki for RAI (Revolt AI Agent)!

RAI

Lightweight, Faster LLM Agents,Team building with YAML Configuration

RAI is a modern, YAML-driven CLI tool and framework for building intelligent agents and agent teams tailored for cybersecurity automation, offensive security, and penetration testing operations. Built on top of the powerful Agno framework, RAI enables security professionals, red teamers, and AI hackers to design, orchestrate, and deploy advanced LLM-powered agents without writing traditional code. Its no-code architecture leverages structured YAML configurations to define agent behavior, tools, and team collaboration logic.

This wiki provides comprehensive documentation to help you get started with RAI, from installation and basic usage to advanced configuration of agents and teams.

Key Features

  • Interactive Shell Mode – Engage in real-time conversations with LLM agents and teams via a powerful interactive CLI. Seamlessly switch between agents or teams with intuitive commands.

  • YAML-Based Agent & Team Building – Define agents and teams using easy-to-edit YAML templates. Accelerate development with low-code configurations and smart defaults.

  • Multi-Agent & Team Support – Build, run, and manage multiple agents or teams in parallel with full operational isolation and coordination.

  • Tool Integration (SSE & stdio) – Integrate custom tools via Server-Sent Events or standard I/O for dynamic agent-tool communication.

  • Dynamic Team Allocation – Flexibly assign, reassign, or reconfigure agents across different teams at runtime to optimize task workflows.

  • MCP-Compatible Infrastructure – Built with modularity in mind, RAI is ready for integration with Model Context Protocol (MCP) tooling and future agent standards.

  • Fast & Flexible Configuration – Lightweight setup with extensible configuration options. Designed for developers who want control without the clutter.

  • Built-In Reasoning Engine – Agents can think, reason, and decide intelligently before taking action, enabling smarter task execution.

  • Cybersecurity-First Design – Purpose-built for red teaming, bug bounty automation, recon, exploit development, and offensive security workflows.

  • Agent-to-Agent Communication – Enable inter-agent messaging within teams, allowing agents to delegate tasks, collaborate, and share results autonomously.

Getting Started

To get started with RAI, follow these steps:

  1. Installation: Install RAI on your local machine using the uv package manager.
  2. Configuration: Set up your first agents and teams using either the GUI tool or manual YAML configuration.
  3. Usage: Learn the basic commands and start interacting with your AI agents.
  4. Examples: Explore practical examples for various cybersecurity scenarios.

RAI in Action RAI interface showing the main application window with agent interaction

Documentation Sections

Core Documentation

Practical Guides

  • Examples: Real-world scenarios and configuration examples
  • Screenshots: Visual examples of RAI in action
  • FAQ: Frequently asked questions and quick answers
  • Troubleshooting: Solutions for common issues

For Developers

Clone this wiki locally