EmbodiedClaw: Conversational Workflow Execution for Embodied AI Development

TL;DR

EmbodiedClaw introduces a conversational system that automates complex embodied AI development workflows, reducing manual effort and enhancing reproducibility across multi-task, multi-scene research settings.

Contribution

The paper presents EmbodiedClaw, a novel conversational agent that automates embodied AI development tasks, streamlining workflow execution and improving research efficiency.

Findings

EmbodiedClaw reduces manual engineering effort in embodied AI workflows.

It improves executability, consistency, and reproducibility of research activities.

Experiments demonstrate effectiveness across various evaluation tasks.

Abstract

Embodied AI research is increasingly moving beyond single-task, single-environment policy learning toward multi-task, multi-scene, and multi-model settings. This shift substantially increases the engineering overhead and development time required for stages such as evaluation environment construction, trajectory collection, model training, and evaluation. To address this challenge, we propose a new paradigm for embodied AI development in which users express goals and constraints through conversation, and the system automatically plans and executes the development workflow. We instantiate this paradigm with EmbodiedClaw, a conversational agent that turns high-frequency, high-cost embodied research activities, including environment creation and revision, benchmark transformation, trajectory synthesis, model evaluation, and asset expansion, into executable skills. Experiments on end-to-end workflow tasks, capability-specific evaluations, human researcher studies, and ablations show that EmbodiedClaw reduces manual engineering effort while improving executability, consistency, and reproducibility. These results suggest a shift from manual toolchains to conversationally executable workflows for embodied AI development.