Choose Rivya image workflows for product photos, ecommerce stills, references, style exploration, Studio iteration, and saved history.

Use this AI image workflow guide before you spend credits on product photos, ecommerce stills, reference-led generation, style exploration, or polish passes in Rivya.

The fastest way to waste image credits is to treat every image request like the same kind of job.

It almost never is.

Image work gets much easier once you separate three things:

what you are actually trying to make
what has to go right first
whether references are central or optional

This page is the workflow reference for image work. If you want the more decision-oriented version about where to begin and how to avoid a weak first run, How to Generate AI Images with Rivya is the better paired guide.

Start With The Job, Not The Model Name

Before you pick a model, decide what kind of image task this really is.

In practice, most requests fall into one of these groups:

a clean product or ecommerce still
a reference-led image system
a more style-led exploratory image
a later polish pass after the direction already works

All of those end in a still image, but they are not the same job.

The Four Image Patterns People Hit Most Often

Right now, the most common image paths look like this:

product-facing, brand, and ecommerce stills often start with Flux 2 Pro
reference-heavy structure and tighter obedience often start with GPT Image 1.5
style-first exploration often begins with Midjourney
cheap first-run direction checks often begin with Z-Image

Those are not hard rules. They are simply the patterns that make the most sense most often in the current lineup.

Start In The Right Place

Use the public image side when you need:

a category overview at /image
a comparison pass in AI Models
a direct model-first entry through /ai-models/[modelSlug]

Move into Studio once you need:

signed-in execution
saved continuity
reference uploads
repeated work on the same direction

That is why a lot of image work starts publicly but becomes real in /studio/image/[modelSlug].

References Change The Decision Early

Image work becomes easier the moment you decide whether references are central or optional.

If the task depends on:

one anchor image
a handful of references
a controlled multi-image system

that should affect the model choice before the first run, not after a few wasted attempts.

That is also why References and Uploads in Rivya matters so much here.

What A Good First Image Run Looks Like

A strong first image run usually looks like this:

decide what the real deliverable is
compare one or two likely models
decide whether references are part of the job
sign in before upload or execution if the workflow needs account context
write the prompt for the actual asset, not just a loose visual idea
review the result before changing the whole setup

That sequence cuts down a lot of avoidable reruns.

History Is Part Of The Workflow

Image work gets more useful once you stop treating each run like a disposable card.

History matters because it lets you:

reopen a result later
compare nearby attempts
use a strong still as the basis for video work
continue the same project instead of starting over

That is one reason image work in Rivya feels more like a workflow than a one-shot generator.

Common Image Mistakes

The most common mistakes are:

choosing by brand name before checking the job
realizing too late that references were central
paying for a polish pass before the direction is proven
abandoning a useful result instead of continuing from history

Most of these are workflow mistakes, not model mistakes.

Image Workflow Checklist

When the output should be a still visual, check:

Decide whether the first run is discovery, product delivery, reference editing, or finishing.
Check whether references are required before choosing the model.
Match aspect ratio and resolution to the final placement before spending credits.
Use cheaper direction checks before high-finish passes when the brief is unproven.
Review product accuracy, text, artifacts, and brand fit before making variants.

Recheck Before Increasing Quality

Recheck when the first image fails because of prompt ambiguity, reference role, crop, or model fit. Raising quality rarely fixes a brief that is pointed at the wrong image job.

Rivya AI Image Workflow Guide