Upload reference
Bring in a new image from your device when the next pass needs a specific face, product, sketch, room, or material sample.
Reference Control
Bring the exact images you want the next pass to follow.
Reference control in just4o means you can build the next image from uploads, past generations, just4o Files, and saved Wikipedia tags, then keep those inputs visible while you write the next prompt.
The selected model decides the total image-input budget, and the active source image already consumes one slot when there is one.
Use referenceWhat stays visible
The reference set, the source image, and the next prompt on the same screen.
What the feature is for
Reference control keeps the visual brief next to the prompt.
This feature matters whenever the next image should follow something specific: the same subject, the same product, the same material language, the same lighting idea, or the same research source. Words alone are often not stable enough for that kind of direction.
In just4o, the useful part is not merely attaching an image. It is that the reference set stays beside the prompt while you write, so the images you are directing from are still in view at the exact moment the next request is shaped.
That is what makes reference control practical for real work rather than a hidden attachment step.
In the editor today
Four sources, one input budget.
The attach flow already pulls from uploads, past generations, just4o Files, and saved Wikipedia tags, and the active source image is counted against the model limit before extra references are sent.
Generations and just4o Files
Reuse a result you already made or pull a long-lived asset back in from Files when the work needs continuity across multiple passes.
Saved Wikipedia tags
Tag useful article images into your account and reuse them later as grounded visual research instead of one-off browser inspiration.
1 total image input
FLUX Kontext routes are configured for tightly anchored image-to-image work.
FLUX.1 Kontext Pro · FLUX.1 Kontext Max
3 total image inputs
Most image models in just4o live here, including the source image when one is active.
Nano Banana / Gemini 2.5 Flash Image · Imagen 4 Fast · Imagen 4 · Imagen 4 Ultra · Grok Imagine · Grok Imagine Pro
8 total image inputs
The biggest reference boards live on the routes configured for larger multi-image blends.
ChatGPT Image Latest · Nano Banana Pro / Gemini 3 Pro Image · Nano Banana 2 / Gemini 3.1 Flash Image
Open Images and build the next pass from the right reference set.
Upload a new source, reuse a generation, pull from Files, or save a Wikipedia image, then keep those inputs beside the prompt while you direct the result.