Granufield

Summary

Granufield is a web app that generates Lightroom presets from a reference image. It is the next phase of the preset packs I already sold.

• Problem: with a ready-made pack, people like one or two presets and rarely all of them.
• What I did: before building, I surveyed the pack's customers. Then I created the product from the code to the billing, with an engine that reads the reference and builds the preset file.
• Result: the product is ready and in production at granu.app. It has not been put up for sale yet, because the engine is in its final adjustments.
• Stack: an AI vision model on the server, Lightroom .xmp files and Hotmart.
• Background: Granu Presets' own product, created in 2026.

Context

I like to give a touch of visual identity to everything I create.

I always created my own presets, the filters I applied to my Instagram posts and stories.

After a while, many people started asking what filter that was.

That was when I turned my personal presets into a pack and put it on the market as a digital product.

I called that pack GRANU Presets Pack. The word GRANU came from the grainy effects that were always in my presets.

It always sold well.

But a ready-made pack has a limit. A person likes one preset or another and hardly all of them.

As my knowledge evolves, I always try to modernize and simplify what I already do and what I sell.

I thought: "What if I built something that gave people the autonomy to create their own pack, with styles 100% to their taste?"

After the idea, instead of just starting to build, I ran a survey with all GRANU Presets customers to find out whether a solution like this would be welcome.

The result was very positive. Several customers answered that they would be interested.

What I did

I created Granufield, a web app that generates Lightroom presets from a reference image.

The person brings the reference aesthetics, and the tool generates the presets for them.

I took care of the whole build, from the code to the billing.

How it works

 

The full path has six steps, from the purchase on Hotmart to the preset applied in Lightroom:

The product economy

In the end, what Granufield sells is tokens. Inside the app I call them credits.

Each preset generation uses one credit.

Credits come in two ways:

• in the plan, which brings a quota of credits per month;
• in one-off credits, for those who used the month's quota and want to continue without changing plans.

The plan quota renews every month and does not accumulate. One-off credits do not expire.

The engine

The engine is the heart of the product. It is what analyzes the reference image and turns what it saw into a preset.

Generating a preset is not generating an image. It is analyzing the reference and translating its aesthetic into Lightroom parameters.

That is why the engine uses a vision model, which reads images, and not an image generator.

It runs entirely on the server. The AI key never reaches the browser, and the server is where the account and the credit are checked before each generation.

The AI does not write the file. It returns the parameters, the code validates each value and builds the .xmp, the Lightroom preset format.

This way the file the person downloads is always a valid preset.

Result

The result is a product that is ready and functional, in production at granu.app.

It has not gone to market yet because I am making the last adjustments and tests on the engine, so that it delivers presets as close as possible to the reference images.

Granufield is a new phase, the natural evolution of selling only preset packs.