Glutamate Guard began with a small but persistent shopping problem: ingredient lists are dense, terminology varies, and checking every label by memory is slow. The resulting app compresses that task into a private camera workflow on an iPhone.
An iOS app that photographs an ingredient list, reads it with on-device OCR, checks the text against a curated term list, and returns a simple reading.
The aim. Turn a difficult ingredient-list check into a quick, private phone workflow.
01The problem behind the project
It was built in response to a real need for a fast ingredient check when no suitable App Store option was available.
The narrowness of the problem shaped the product. It did not need accounts, a social layer, a remote database, or a generalized nutrition engine. It needed to capture text, compare it with a curated list, and explain that result without implying medical certainty.
People who choose to monitor glutamate-related ingredient terms. The app is not a medical device and cannot determine individual sensitivity or safety.
02How it took shape
SwiftUI and Apple’s Vision framework. The workflow runs offline with no backend, no ads, and no collection of ingredient images.
Apple's Vision framework performs the OCR on-device, while SwiftUI supplies the capture and result flow. Keeping the entire check offline reduced both latency and privacy exposure: ingredient photographs do not need to leave the phone merely to answer a term-matching question.
Josiah chose the problem, designed the minimal workflow, built and tested the app with AI assistance, and submitted it to the App Store.
Version 1.1 cleared App Store review. A working MVP ran on a real device and was submitted the same day it was built.
03What the project means now
Its significance is partly methodological. A one-purpose health-adjacent tool can be useful precisely because it refuses to become a diagnostic product. The remaining work is therefore not feature expansion so much as clearer documentation of the term list, evidence boundary, and visual presentation.
The term-list methodology needs clearer public documentation, and the product must retain a non-medical disclaimer without overstating the science of sensitivity.
A narrow problem, a real user, and a short path to handoff can matter more than feature breadth.
Document the term list, current listing state, usage evidence, and public-safe product story.