Black Forest Labs' editing model, and the default behind Image to Image and the seven redraw tools here. It is the quickest way to change one thing in a photo and leave everything else alone.
Sign in to start — new accounts get 20 free credits, no card.
The same prompts sit above the input box; click one there to load it.
Matte black ceramic mug on a pale linen surface, soft window light from the left, shallow depth of field, editorial product photography
Close-up portrait of an older woman laughing, natural window light, film grain, 85mm lens, shallow depth of field
Misty pine forest at dawn, low fog between the trunks, cold blue light, wide shot
Give it a photo and an instruction. It is the model the site reaches for whenever the job is "change this part, keep the rest", which is most of the toolkit.
Our edit runs came back in 18 to 27 seconds. Nano Banana, at the same price, can take several minutes on the same job when it is busy.
Its input field takes a single image rather than a list. That is why Virtual Try-On does not offer it: it would see the person and have to guess the garment.
Speed here does not cost extra. Spend the difference on running two variations instead of one.
Pick any of these from the Model menu above; the price is shown before you generate. Each name links to what that model is good at. Compare every tier.
Google Gemini 2.5 Flash · fastest, cheapest
5 credits per image
Black Forest Labs · fastest, about 5s per image
5 credits per image
ByteDance · good at Chinese text and posters
7 credits per image
ByteDance · most photorealistic, highest detail
18 credits per image
OpenAI · good with text in images, half the price of 2.5
7 credits per image
OpenAI · for print or heavy cropping
10 credits per image
OpenAI · newest, better instruction following
13 credits per image
OpenAI · newest, highest resolution
20 credits per image
JPG, PNG, WebP or HEIC up to 10 MB. Phone photos are fine.
"Make it winter, keep the building and the people exactly as they are." One change per run beats a list of five.
Each run is cheap and quick, so stack edits across runs instead of writing one long prompt.
It is fast and cheap enough that chaining beats stacking. Each pass gives you something to check.
"Keep the face, pose and background" is the single most useful sentence you can add to an edit prompt.
"The wall is painted sage green" works better than "repaint the wall".
If you add an object, say where the light comes from so it sits in the scene instead of floating on top.
Start free. Every tool draws from one balance — no separate subscriptions.
$0 / month
Enough to see whether the results are any good
$6.9 / month
Enough to find out if this works for you
$13.9 / month
For people editing every week
$39 / month
For teams and heavy output
An image editing model from Black Forest Labs, the lab behind the FLUX family. It takes a photo plus an instruction and returns the photo with that one change applied.
Both cost 5 credits. Kontext is considerably faster on edits, 18 to 27 seconds against several minutes in our runs. Nano Banana follows long instructions more literally and can read several reference images; Kontext takes one.
Yes, it is offered in Text to Image as well. Its strength is editing, but it will generate from a prompt.
That tool needs a model that reads two images, the person and the garment. Kontext takes a single image, so it would never see the garment and would invent one instead. It is left out of that dropdown on purpose.
On single-subject edits, yes, especially when you tell it to. Add "keep the face exactly the same" and check the result at full size.
You are only charged when an image comes back. A failed or timed-out run costs nothing.
Kontext is preselected in the generator above. Upload a photo and name one change.
All of them are in the same dropdown, and one credit balance covers every one.