Welcome to RPent#
RPent (Recursive Physical Agent) is an open framework for building embodied agents that continuously evolve through recursive interaction with the physical world. Rather than prescribing a single foundation model, RPent provides a recursive agent framework that harnesses heterogeneous intelligence, including perception, reasoning, memory, execution, and self-evolution, into a unified physical agent. Through continuous interaction, reflection, and adaptation, RPent enables physical agents to acquire new capabilities and evolve beyond their initial design.
What RPent is, what the pentagram + ∞ logo means, and the high-level architecture at a glance.
Clone RPent and install the whole stack with a single
pip install.
Configure an LLM API key and the VLA checkpoint, then run one LIBERO task end-to-end.
Drive the LIBERO / RoboCasa simulators or a Franka / SO-101 arm, switch planners, and pick action primitives.
RPent’s implementation-level architecture, plus how to add a new robot, a new action primitive, or extend memory.