Active Beta — v1.3.1

Helping Detect
Cataracts Earlier —
We Need Your Help

iSpeye is an AI-powered cataract screening app currently in active beta testing. We're a small team building something we believe can make a difference — and your feedback as a tester is shaping what it becomes.

Beta
v1.3.1 — Active Testing
Dual AI
ResNet50 + DenseNet121
Free
For Approved Testers

Cataracts Are the Leading Cause of Blindness

Cataracts account for over 51% of world blindness — affecting an estimated 94 million people globally. In Sub-Saharan Africa, limited access to ophthalmologists means most cases go undetected until vision loss is severe.

The good news: 90% of visual impairment from cataracts is preventable with early detection and timely treatment. We believe a smartphone can be the first line of defense.

51%
of world blindness
94M
people affected
90%
preventable

We're Still Building — And Testing

iSpeye is in active beta. The core AI works — it uses a ResNet50 and DenseNet121 ensemble to analyze eye photographs for cataract indicators — but we're still refining accuracy, improving the user experience, and ironing out bugs.

As a beta tester, you get early access in exchange for honest feedback. Found something broken? Tell us — every report helps.

Core Features in Beta

These features are in the current build. Some may be rough around the edges — that's why you're here.

AI Cataract Analysis

Dual-model ensemble (ResNet50 + DenseNet121) analyzes eye photos and returns a screening result.

Testing accuracy
Photo Capture

Take a photo directly in-app or upload from your gallery. We're testing image quality handling.

In testing
User Accounts

Sign up, log in, and access your screening history. Cloud auth via Supabase.

Working
Screening History

Past screenings are saved to your account. We're testing data persistence and display.

In testing

How iSpeye Works

Three steps — and then we need your honest feedback on all three.

1
Capture

Open iSpeye and take a clear, close-up photo of the eye. Try both camera capture and gallery upload — and let us know which works better.

2
Analyze

The photo is sent to our AI backend. ResNet50 and DenseNet121 both process it and their predictions are combined. Note the response time — report if it's slow.

3
Review & Report

View your result. Does it make sense? Is the result clear? Then report what you found — good or bad.

Download the Beta Build

Access is limited to approved testers. Enter your email — if you're on the list, you'll go straight to the download. Otherwise you can apply to join the test group.

Not on the list? Request to Join Beta

Already testing the app?

Report a Bug
Android 8.0+
~80 MB
Internet required