Read everyday text
Ask for help with mail, labels, menus, signs, displays, and controls visible in a selected image.
Assistive AI · Working prototype
Understand more of the world, on your terms.
Access AI is designed around blind and low-vision users. Ask with text or voice, or choose an image task when you want help reading, identifying, or describing something.
Practical use cases
Help with the task in front of you
The prototype focuses on bounded, user-requested assistance. Results can be incomplete or wrong, so important information should be checked with a trusted source.
Identify objects
Request a practical description of an object, its visible features, and uncertainty when identification is unclear.
Describe a scene
Get a concise account of visible people, objects, relative positions, and useful nonvisual landmarks.
Ask by typing
Use a standard text field with a software keyboard, hardware keyboard, or compatible Braille input service.
Ask by voice
Record a question when voice is more convenient, then receive text and an optional generated spoken answer.
Choose when to use an image
Image analysis begins only after you select or capture an image and activate the related task.
Built from lived experience
A product direction informed by vision loss
Founder Anthony Pope's personal experience with vision loss informs the emphasis on nonvisual use, direct control, clear feedback, and honest limits. The project is owned by Palmetto Ridge Holdings LLC, operating through Palmetto Ridge Media Group.
Clear product status
Working, in development, and planned
Working
Browser proof of concept
Text, recorded voice, generated speech, and user-triggered still-image tasks are available in the public Streamlit prototype.
In development
Native Android direction
The current direction prioritizes accessible onboarding, Android semantics, platform speech, and local preferences before production AI features are connected.
Planned
Selective smart-glasses integration
A vendor-neutral design proposes explicit still-frame capture through a phone or controlled relay. No glasses SDK or navigation capability is implemented.
Questions
What to know before trying Access AI
What can I try today?
The working browser prototype supports typed or Braille-keyboard questions, recorded voice questions, generated spoken answers, and user-triggered still-image tasks.
Is Access AI a navigation or safety system?
No. Access AI is development-stage assistive software and is not a navigation, emergency, medical, legal, financial, or safety system.
How can I join the pilot?
Use the pilot page to open a pre-addressed email to the founder. No pilot database or third-party form is used at this stage.
Development-stage safety notice: Access AI can produce incomplete or incorrect answers. Do not rely on it as the sole source for navigation, traffic, obstacles, medication, emergencies, legal or financial obligations, or other safety-critical decisions.