Open Technology, Open Science.
We build open-source technology for autonomous science, to automate complex workflows with human oversight. We integrate techniques such as test time compute scaling, and reinforcement learning within a modular and scalable framework.
Our EVE system orchestrates these methods in a multi-agent, multi-turn, multi-model system. With a human-in-the-loop design, users can steer and intervene in processes, ensuring adaptability and efficiency across diverse scientific applications.
