Vision language models directly control robots via semantic interface
Show-Harness: Just a VLM Agent Can Play Robots
RoboticsArtificial IntelligenceComputer Vision and Pattern RecognitionMultimedia
Summary
Controlling robots using large vision-language AI models is hard because the AI's understanding needs to be turned into robot actions. The researchers created Show-Harness, a simple system that connects what the AI understands with actions the robot can perform. It uses clear, understandable steps so the AI can decide detailed movements, and a special part translates these steps into actual robot commands. They also made a way for humans and AI to teach robots by manipulating computer interfaces without needing special hardware. Their tests show this helps robots perform well in many tasks and settings without extra complicated training.
What this means in practice
- •For robotics developers: Use Show-Harness to enable vision-language models to control various robots without extra robot-specific training.
- •For human-computer interaction designers: Build GUI-based demonstration tools that let humans guide robots and collect training data without specialized hardware.
Authors
Yanzhe Chen, Zechen Bai, Zhijun Cao, Wenzheng Zeng, Kevin Qinghong Lin, Yiqi Lin, Guoqiang Liang, Kevin Yuchen Ma, Qiming Huang, Mike Zheng Shou
Abstract
Foundation vision-language models (VLMs) exhibit broad intelligence about the world, yet translating this intelligence into robot control remains challenging. We present Show-Harness, an Embodied Harness that enables VLMs to "play" robots through a compact semantic interface linking intent to action. Show-Harness exposes discrete semantic action units that VLMs can naturally reason over, while embodiment-specific interpreters deterministically ground them into local robot actions, keeping the VLM directly responsible for fine-grained physical decisions. Through the same interface, Show-Harness demonstrates the feasibility of (1) directly unlocking closed-source frontier VLMs for zero-shot robot control, and (2) adapting small-scale open-source VLMs for low-cost deployment with just a few GPU-hours of fine-tuning. We further develop GUMI (GUI Manipulation Interface), which extends the same semantic action space to GUI-based demonstration collection, allowing humans and agents to "play" robots across embodiments without specialized teleoperation hardware. Extensive experiments show that Show-Harness-equipped VLM agents generalize robustly across tasks, embodiments, and environments, outperforming representative agentic and VLA paradigms. These results suggest that the right interface can unlock substantial embodied capability from foundation VLMs, without requiring additional model capacity or costly embodiment-specific pretraining.