How OLI Works

Hi

again

I’m a small green vegetable with big curiosity.

Every day, I hang out on supermarket shelves, inside pantries, and sometimes in your kitchen. I stare at food labels all day long. Not because I have to, but because I really want to know what’s inside.

I don’t eat food.

I read it.

Every day, I hang out on supermarket
shelves, inside pantries, and sometimes in
your kitchen. I stare at food labels all day
long. Not because I have to, but because I
really want to know what's inside.

Food labels tell stories.
Some are simple.
Some are sneaky.
Some use very fancy words for very
ordinary things.
That's where I come in.

People pick up a box.They turn it around.They squint. They sigh.I hear things like
“What does this mean?”“Is this good or bad?”Why is this ingredient spelled
like that?”

One day, I noticed
something strange

So I thought, what if someone could read labels instantly and explain them in a way that actually makes sense?

That someone became me.

When you scan food,
I wake up The moment you scan
a barcode or point your camera
at an ingredient list, I get
excited.

Boop. I'm awake.
I don't just look at the product
name. I look deeper. I zoom in on the
ingredient list. I capture every word,
even the tiny ones hiding at the
bottom of the package.

You scan.

I read.

This happens using real-time AI text recognition. That means I can read curved labels, odd layouts, small fonts, and messy packaging without you typing a single thing.

I clean up the messy ingredient world

Food labels are not always neat.

Sometimes ingredients are
separated by commas.

Sometimes they hide inside
brackets.

Sometimes they repeat themselves
using different names.

I gently take all that messy text and organize it.

I remove duplicates.

I normalize spellings.

I separate combined ingredients

I keep important codes and numbers that matter.

This step is critical, because bad input creates bad results. I’m very careful here.

Ingredients love disguises. I remove them. Here’s a secret.Ingredients often wear costumes.One ingredient can show up as

E471

INS 471

Mono and diglycerides of fatty acids

No guessing

No internet delay

No human reviewing later.

I recognize ingredients by
what they are, not
what they pretend to be.

Different outfits. Same ingredient.

I use a canonical ingredient system to map all those disguises back to a single known identity. This happens instantly using local logic and verified open food datasets.

Now comes the thinking part

Once I know exactly what ingredients are present, I start thinking.

I look at

Ingredient quality

Additives and preservatives

Artificial colors and flavors

Processing indicators

Ingredient order and prominence

Each ingredient carries meaning. Some bring value. Some raise questions. Some deserve extra attention depending on where they appear in the list.

I don’t panic.

I don’t judge.

I just calculate.

I calculate scores instantly, right on your phone

Here’s something important about me.I don’t wait.Other apps sometimes send scans to a server.Then humans review the data.Then you wait hours or days to see results.I don’t do that.

My scoring logic runs automatically
and deterministically in real time.
The moment scanning finishes, the
score appears.

That means / No delays
No manual intervention
No inconsistency
The same product always gets the same result.

OLI doesn't stop at one scan.
Over time, patterns start to appear.
You notice which ingredients show up often.
You recognize certain additives.
You understand labels faster than before.

That's the real goal.

Not control. | Not rules. | Just understanding.