Fig. 2
From: Large language models open new way of AI-assisted molecule design for chemists
System architecture of ChatChemTS. LangChain was used to implement the ReAct framework in ChatChemTS. LLMs were used in the agent, the reward generator tool, and the configuration generator tool. Additionally, both tools utilized LLMs specialized in the corresponding tasks using in-context learning (ICL). ChatChemTS was containerized using Docker, allowing for easy deployment to user environments through Docker Compose. Users can employ ChatChemTS via local web browsers, such as Google Chrome