Stretch for AI

An Open Platform for Embodied Intelligence

Open Models, Data, and Code

The power of the Stretch community

Stretch has a vibrant community of Embodied AI researchers sharing their models, data, and code, including popular projects such as ForceSight, Dobb-E, HomeRobot OVMM, RT-X, Teach A Robot to Fish and VRB.


Real World Data Capture

Take Stretch out of the lab

Portability matters. Don’t let cumbersome robots slow you down.

The OK-Robot project took Stretch to 10 NYC homes to accomplish 171 tasks.

OK-Robot : An open, modular framework for zero-shot, language conditioned pick-and-drop tasks in arbitrary homes : ok-robot.github.io

Python Centric

The beauty of Python at your fingertips

ROS isn’t required to use Stretch. Our Python SDK and autonomy demos increasingly support developers working in pure Python.

Flexible Compute

On-board and off-board solutions

The on-board Intel NUC 12 i5 is powerful enough to run AI models such as Dobb-E and YOLOv8.

Off-board GPUs are widely used by the community for real-time control and supported by tutorials and the Stretch WiFi Access Point.

Examples from the Community

  • OK-Robot

    NYU, AI at Meta

    An open, modular, framework for zero-shot language conditioned pick-and-drop tasks in arbitrary homes.

  • Dobb-E

    NYU, Meta

    An open-source, general framework for learning household robotic manipulation.

  • SPOC

    Allen Institute for AI

    Imitating Shortest Paths in Simulation Enables Effective Navigation and Manipulation in the Real World.

  • HomeRobot

    Georgia Tech, Meta AI, Carnegie Mellon, Simon Fraser

    HomeRobot: Open-Vocabulary Mobile Manipulation (OVMM)

  • WHIRL

    Carnegie Mellon University

    Human-to-Robot Imitation in the Wild

  • ForceSight

    Georgia Tech

    Text-Guided Mobile Manipulation with Visual-Force Goals

  • SLAP

    Carnegie Mellon, Meta AI, SUNY Binghamton

    Spatial-Language Attention Policies

  • RT-X

    Community Effort

    Open X-Embodiment: Robotic Learning Datasets and RT-X Models

  • VRB

    Carnegie Mellon University, Meta AI

    Affordances from Human Videos as a Versatile Representation for Robotics

Still have questions?

Want to learn more — or see a demonstration of Stretch in action?

Drop us an email or fill out the form below. A member of our team will get back to you right away!