Eyedaptic launches Eva, an agentic AI visual assistant for patients with retinal disease

Now available on the EYE7 smart glasses platform, Eva is designed to go beyond answering questions about the visual environment by reasoning about user needs and directly controlling the device’s vision-enhancement features.

Eyedaptic launches Eva, an AI visual assistant for patients with retinal disease

Eyedaptic, a developer of vision-enhancement technology, has announced the launch of Eva, an agentic AI visual assistant integrated into its EYE7 Vision Enhancing Smart Glasses. According to the company, the release marks a move away from passive, question-and-answer AI toward a proactive agent that supports users in real time.

Eva pairs agentic AI with Eyedaptic’s computer vision platform so that the AI and the smart glasses work as a single system to help users see, understand, and interact with their surroundings. According to the company, Eva demonstrated greater than 99% reading accuracy and 97% command accuracy in internal testing.

From visual assistant to visual agent

Eyedaptic has been adding AI to its wearable platform for several years. In August 2024, the company unveiled the EYE6 with Ivy, a wearable vision aid that paired vision enhancement with an AI-powered visual assistant for patients with age-related macular degeneration (AMD) and other retinal disorders. The EYE7 followed in March 2026, launched alongside the Eyedaptic Visual Assistant Family. At that launch, Ivy was fully available, while Eva was offered through an Early Access Beta Program.

Now Eva is available on the EYE7 smart glasses platform

With the full launch, Eyedaptic describes Eva as a new category of assistive technology because it brings three capabilities into one system: visual intelligence, reasoning based on agentic AI, and direct control of the glasses’ vision-enhancement functions. The company has branded this architecture Eyegentic AI.

The practical difference lies in what the assistant can do with what it sees. Earlier visual assistants mostly respond to a user’s question about a scene, such as reading a sign or identifying an object. Eyedaptic says Eva is built to infer what the user needs, reason about it, and then act on it. That could include adjusting the glasses’ enhancement settings without the user having to work through menus or a remote. The company argues this makes wearable vision assistance more intuitive and responsive.

“Eva is more than an assistant; it’s an intelligent partner,” said Jay Cormier, CEO and founder of Eyedaptic, in the release. The assistant’s name is also personal: Cormier noted that Eva was the name of his late grandmother, whose AMD inspired him to found the company.

Built on Simulated Natural Vision

Eva runs on top of Eyedaptic’s proprietary Simulated Natural Vision software, which is designed to go beyond conventional magnification to help patients with conditions such as AMD regain function in daily life. Eyedaptic says its approach is protected by a portfolio of granted and pending patents, and it describes its platform as the only solution that integrates both advanced vision enhancement and agentic AI.

Independent data on the underlying hardware have recently started to appear. At the American Retina Forum annual meeting in August 2026, David Almeida, MD, MBA, PhD, FRCSC, presented an unsponsored real-world evaluation of the EYE6 in 19 patients with advanced retinal disease. Seventeen of the 19 patients (89%) showed measurable improvement in near visual acuity immediately after fitting, with a mean gain of 5.7 lines.

Context for clinicians

Wearable low vision technology has developed quickly. Modern Retina’s look back at 2024 highlighted augmented reality (AR) headsets, text-to-speech, voice activation, and AI features in many commercially available low vision devices. Earlier platforms, such as eSight Go, focused mainly on enhancing the image a patient sees. Meanwhile, mainstream consumer AI glasses are increasingly being dispensed as low vision devices in rehabilitation settings.

Agentic control is the next step: an assistant that can change how the device works, not just describe the world. For patients with geographic atrophy or advanced neovascular AMD, who often have limited treatment options for restoring function, hands-free, context-aware assistance could make these devices easier to use day to day.

Clinicians should keep the evidence base in view, though. The reading and command accuracy figures come from the company’s own testing and have not been independently validated or peer reviewed. As with AI tools elsewhere in retina care, real-world validation across varied patient populations will determine how well Eva performs outside controlled conditions. Patient selection, training, and realistic expectations also remain central to successful low vision rehabilitation.

Eva is available now to all users of the EYE7 platform and can be installed through a remote software upgrade, without new hardware.