code-act

CodeAct is a system that enhances large language model (LLM) agents by enabling executable code. It integrates a Python interpreter to perform code execution and supports multi-turn interactions for dynamically correcting previous actions or issuing new actions based on new observations. CodeAct aims to improve the success rate of LLM agents when handling complex tasks. Users can deploy and use the CodeActAgent in multiple ways, including Docker and Kubernetes. The system is suitable for developers and researchers who need efficient and flexible code execution capabilities.