This team is developing a large-scale bi-directional neural interface that will restore high-fidelity vision to people blinded by retinal degeneration. As a key step in this technology development, they propose here to design, build and test in vivo the first device suitable for acute use in humans. The device will incorporate the high-resolution recording and stimulation chip they have recently fabricated, recording and stimulation algorithms they have developed and tested in a laboratory prototype, and surgical devices and approaches we have developed and tested in macaques. The assembled device is designed to be introduced into the eye of an awake human patient for ~1 hour during a retinal surgery, with retinal recording and stimulation performed to calibrate the device to the patient, and then stimulation provided to test the quality of the evoked visual sensations. The key steps in developing the device involve electronic packaging and encapsulation, surgical placement, and optimizing recording and stimulation algorithms for real-time use in vivo. By the end of one year, they intend to have completed the device fabrication and performed the first test of its function in macaques. This will pave the way for first-in-human experiments, the immediate next step and the key proof of concept for commercialization.