MotleyAI/motleycrew

What it solves

Motleycrew addresses the complexity of building multi-agent AI systems by providing a unified framework to orchestrate agents and tools from disparate ecosystems. It allows developers to combine components from frameworks like Langchain, LlamaIndex, CrewAI, and Autogen into a single, cohesive workflow.

How it works

The framework acts as a conductor for AI agents. It uses a knowledge graph to manage task dependencies and data flow, allowing you to define complex interactions between agents. You can chain tasks using simple operators (like >>), and all components are compatible with Langchain's Runnable API. The system also includes built-in support for caching HTTP and LLM calls, as well as observability via integration with platforms like Lunary.

Who it’s for

This is for developers building sophisticated, multi-agent AI applications who want to leverage the best tools from multiple existing frameworks without being locked into a single ecosystem.

Highlights

• Cross-framework integration: Use agents and tools from Langchain, LlamaIndex, CrewAI, and Autogen.
• Knowledge graph backend: Tasks and data are managed via a graph, enabling complex, non-linear workflows.
• LCEL compatibility: All components implement the Langchain Runnable API.
• Built-in observability and caching: Includes native support for Lunary and HTTP request caching to simplify debugging and production monitoring.