OPEN TEACH is an open-source teleoperation system for robotic manipulation that supports multiple arms and hands, allows mobile manipulation, is calibration-free, and works across both simulation and real-world environments. It uses a VR headset for teleoperation, offering low latency and high-frequency visual feedback, enabling real-time correction of robot errors. The system allows users to perform a wide range of tasks, including making a sandwich, ironing cloth, placing items in a basket, lifting it, approaching a cabinet, and opening it. OPEN TEACH is fully open-source and compatible with various robotic platforms, including Franka, xArm, Jaco, Allegro, and Hello Stretch. It supports tasks across different robot morphologies and has been tested on 38 tasks in both simulation and the real world. The system's data can be used to train effective, general-purpose manipulation behaviors, and user studies show significant improvements in teleoperation capability compared to existing frameworks. OPEN TEACH is designed to be user-friendly, with a simple setup and high-frequency visual feedback, making it suitable for both experienced and new users. The system is also flexible, allowing for easy integration with various robotic setups and providing real-time streaming and low-latency visual feedback. OPEN TEACH is an open-source framework that promotes broader adoption and encourages contributions from the robotics community.OPEN TEACH is an open-source teleoperation system for robotic manipulation that supports multiple arms and hands, allows mobile manipulation, is calibration-free, and works across both simulation and real-world environments. It uses a VR headset for teleoperation, offering low latency and high-frequency visual feedback, enabling real-time correction of robot errors. The system allows users to perform a wide range of tasks, including making a sandwich, ironing cloth, placing items in a basket, lifting it, approaching a cabinet, and opening it. OPEN TEACH is fully open-source and compatible with various robotic platforms, including Franka, xArm, Jaco, Allegro, and Hello Stretch. It supports tasks across different robot morphologies and has been tested on 38 tasks in both simulation and the real world. The system's data can be used to train effective, general-purpose manipulation behaviors, and user studies show significant improvements in teleoperation capability compared to existing frameworks. OPEN TEACH is designed to be user-friendly, with a simple setup and high-frequency visual feedback, making it suitable for both experienced and new users. The system is also flexible, allowing for easy integration with various robotic setups and providing real-time streaming and low-latency visual feedback. OPEN TEACH is an open-source framework that promotes broader adoption and encourages contributions from the robotics community.