Rivet is an open-source visual programming environment developed by Ironclad Research for building AI agents using Large Language Models (LLMs). It enables teams to design, debug, and collaborate on complex LLM prompt graphs visually. Rivet allows users to create intricate chains for AI agents efficiently, providing a unique combination of a visual interface, easy debugging capabilities, and remote execution. Users can iterate on prompt graphs within Rivet and seamlessly integrate them into their applications. The tool offers real-time observation of prompt chain execution, enabling developers to debug remotely. Rivet graphs are stored as YAML files, facilitating versioning and collaboration within teams. The tool has received positive feedback from the community for its visual nature, collaborative features, and ability to streamline the development of AI-powered applications. Rivet is utilized by Ironclad Research and has been instrumental in creating products like Ironclad Contract AI, a virtual contract assistant powered by AI agents.