Xaibo
The Modular AI Agent Framework
Build robust, production-ready AI agents that are easy to test and maintain. Stop fighting complexity, start shipping with clean, protocol-based design.
What is Xaibo?¶
Xaibo is a powerful, protocol-driven framework that enables developers to build sophisticated AI agents with unprecedented flexibility and modularity. By using well-defined interfaces and dependency injection, Xaibo allows you to create, test, and deploy AI systems that are both robust and easily maintainable.
Quick Start
Prerequisites: Python 3.10 or higher installed
Get up and running with Xaibo in 60 seconds:
This will give you an agent that you can start modifiying immediately.Follow our Tutorial to learn how everything works.
Why Choose Xaibo?¶
🧩 Swap Models & Tools Instantly
A modular architecture saves time and reduces complexity. Want to switch from OpenAI to Anthropic? Just change a single line in your configuration.
🔌 Test with Confidence
Components communicate through well-defined protocols. This creates clean boundaries, enabling superior unit and integration testing with predictable mocks.
🔍 Visualize Every Step
Transparent proxies capture every component interaction, providing detailed runtime insights and a powerful visual debugger so you always know what your agent is doing.
See Xaibo in Action: The Visual Debug Interface¶
Xaibo includes a powerful debug UI that visualizes your agent's operations in real-time:

Sequence Diagram Overview

Detail View of Component Interactions
A Framework Designed for Developers¶
-
Protocol-Based Architecture
Components interact through well-defined interfaces, enabling easy testing with mocks and ensuring clean boundaries within your system.
-
Dependency Injection
Explicitly declare component dependencies. This makes it trivial to swap implementations - for testing or for production - without refactoring your code.
-
Transparent Proxies
Every component is automatically wrapped with observability proxies that capture parameters, timing, and exceptions for complete visibility.
-
Comprehensive Event System
A built-in event system provides real-time monitoring, call sequence tracking, and performance insights out of the box.
-
Arbitrary Tool Support
Use tools implemented in simple python, use MCP servers or integrate whatever other APIs you want as tools.
-
Production-Ready API Server
Includes a built-in web server with an OpenAI-compatible API and MCP (Model Context Protocol) support for seamless integration into existing ecosystems.
Your Gateway to the Xaibo Ecosystem¶
-
Your first step. Build a complete AI agent with tools and understand Xaibo's core architecture.
-
Practical recipes for installation, tool integration, LLM configuration, and deployment.
-
A deep dive into protocols, modules, dependency injection, and Xaibo's design principles.
-
Complete technical documentation for every module, protocol, and CLI command.
-
Learn to create custom Python and MCP tools that extend your agent's capabilities
-
Understand Xaibo's protocol-based architecture and transparent proxy system
Join the Community¶
-
Contribute to the source code, file issues, and track development progress.
-
Join our community for support, to ask questions, and to share what you're building.
-
Get in touch for direct inquiries and support.
Ready to Build Better Agents?¶
Start the Tutorial to create your first Xaibo agent, or dive into Core Concepts to understand the framework's architecture.