Agent framework driven by knowledge graph engine
Top 46.4% on sourcepulse
muAgent is an LLM-driven agent framework designed for complex Standard Operating Procedure (SOP) automation, particularly in DevOps scenarios. It targets developers and researchers seeking to orchestrate multi-agent systems, leveraging knowledge graphs (KG) for enhanced reasoning, collaboration, and human interaction. The framework aims to simplify the creation of intelligent agents through a visual, drag-and-drop interface and SDK.
How It Works
muAgent utilizes an Eventic Knowledge Graph (EKG) as its core engine, integrating LLMs with structured knowledge. This approach allows for complex reasoning by combining LLM flexibility with KG's deterministic structure, enabling human guidance and knowledge-on-demand capabilities. Key features include EKG Builder for designing agent interactions, EKG Assets for schema-driven automation, and EKG Reasoning for flexible, human-guided decision-making.
Quick Start & Requirements
docker network create ekg-net
), and start services with docker compose up -d
. Access the UI at https://localhost:8000
.pip install codefuse-muagent
.Highlighted Details
Maintenance & Community
The project is open-source as of April 2024. Contribution is encouraged via GitHub Issues.
Licensing & Compatibility
The repository does not explicitly state a license in the README.
Limitations & Caveats
The muagent-related Docker images are currently in beta. The README does not specify compatibility with commercial or closed-source applications.
6 months ago
1 week