UniManip: General-Purpose Zero-Shot Robotic Manipulation with Agentic Operational Graph

TL;DR

UniManip introduces a unified framework combining semantic reasoning and physical grounding through an agentic operational graph, enabling robust zero-shot robotic manipulation in unstructured environments with high success rates.

Contribution

The paper presents UniManip, a novel bi-level agentic operational graph framework that unifies semantic reasoning and physical grounding for zero-shot robotic manipulation.

Findings

Achieves 22.5% higher success rate than state-of-the-art VLA models.

Demonstrates 25.0% higher success rate than hierarchical baselines.

Enables zero-shot transfer from fixed-base to mobile manipulation without reconfiguration.

Abstract

Achieving general-purpose robotic manipulation requires robots to seamlessly bridge high-level semantic intent with low-level physical interaction in unstructured environments. However, existing approaches falter in zero-shot generalization: end-to-end Vision-Language-Action (VLA) models often lack the precision required for long-horizon tasks, while traditional hierarchical planners suffer from semantic rigidity when facing open-world variations. To address this, we present UniManip, a framework grounded in a Bi-level Agentic Operational Graph (AOG) that unifies semantic reasoning and physical grounding. By coupling a high-level Agentic Layer for task orchestration with a low-level Scene Layer for dynamic state representation, the system continuously aligns abstract planning with geometric constraints, enabling robust zero-shot execution. Unlike static pipelines, UniManip operates as a dynamic agentic loop: it actively instantiates object-centric scene graphs from unstructured perception, parameterizes these representations into collision-free trajectories via a safety-aware local planner, and exploits structured memory to autonomously diagnose and recover from execution failures. Extensive experiments validate the system's robust zero-shot capability on unseen objects and tasks, demonstrating a 22.5% and 25.0% higher success rate compared to state-of-the-art VLA and hierarchical baselines, respectively. Notably, the system enables direct zero-shot transfer from fixed-base setups to mobile manipulation without fine-tuning or reconfiguration. Our open-source project page can be found at https://henryhcliu.github.io/unimanip.