OpenAGI

OpenAGI: When LLM Meets Domain Experts

open-sourceagent-frameworks
Visit WebsiteView on GitHub
2.3k
Stars
+188
Stars/month
0
Releases (6m)

Overview

OpenAGI is a Python package for creating AI agents designed to work with AIOS (AI Operating System). It provides a structured framework for building domain expert agents that can leverage external tools and integrate with larger AI ecosystems. The package follows a research-backed approach from the paper 'OpenAGI: When LLM Meets Domain Experts' published in NeurIPS 2023. Developers can create agents with specific folder structures, integrate external tools, and share their agents with the community through upload/download functionality. The system emphasizes modularity and reusability, allowing agents to be easily distributed and deployed. While originally designed as a standalone agent creation tool, OpenAGI now serves as the foundation for building agents that integrate with AIOS through the Cerebrum SDK. The package includes built-in support for external tools and provides a standardized way to package agent logic, configurations, and dependencies together.

Pros

  • + Research-backed framework with peer-reviewed methodology published in NeurIPS 2023
  • + Structured agent sharing ecosystem with upload/download functionality for community collaboration
  • + Built-in external tool integration system allowing agents to leverage specialized capabilities

Cons

  • - Requires migration to Cerebrum SDK for full AIOS integration, suggesting the main package may have limited standalone utility
  • - Rigid folder structure requirements that may limit flexibility in agent organization
  • - Heavy dependency on AIOS ecosystem for optimal functionality

Use Cases

Getting Started

1. Install via pip: `pip install pyopenagi` 2. Create agent folder structure under pyopenagi/agents/author/agent_name with agent.py, config.json, and meta_requirements.txt 3. Implement agent logic and test locally, then optionally upload to share with community