Black Forest Labs' in-context editing model changes exactly what you ask for and leaves the rest of the picture alone. Same character, same style, new scene. Kontext Pro and Kontext Max, 2 or 3 credits per edit, no watermark. Available on every Aura AI plan.
FLUX Kontext, officially FLUX.1 Kontext, is the editing branch of the FLUX family from Black Forest Labs. It launched on May 29, 2025 in two hosted versions, Kontext [pro] and Kontext [max], with an open-weight [dev] variant in private beta at the time. The idea behind it is "in-context" generation: the model reads an image and a text instruction together, so instead of describing a whole new picture you describe only the change. Change the outfit, move the subject to another city, swap the time of day or rewrite the words on a sign, and everything you did not mention stays as it was.
That is what separates it from the original FLUX 1 Dev, which only goes from text to image. Kontext preserves the identity of a character across several rounds of edits, keeps the style of a reference consistent, and handles local edits without repainting the full frame. Black Forest Labs reported inference up to 8 times faster than FLUX 1.1 Pro at launch, together with top scores on its own KontextBench for text editing and character preservation. Pro is tuned for iterative multi-step refinement; Max pushes prompt adherence and typography further at a higher price.
On Aura AI, FLUX Kontext is an editing model only. You find FLUX Kontext Pro (2 credits) and FLUX Kontext Max (3 credits) in the AI image editor next to FLUX 2 Edit, Nano Banana Edit and GPT Image 2 Edit. Upload any image, or send a result from the text to image tool, and get the edited version back without watermark.
Six things Kontext does that a text to image model cannot
Kontext takes your picture and your instruction as one input. The prompt describes the change, not the whole scene, so a two-line request is usually enough to get a precise result.
The same face, body and outfit survive across edits. Put a character in a new location, a new pose or a new lighting setup and they still read as the same person.
Change one object, one colour or one region and leave everything else untouched. Kontext does not re-render the unrelated parts of the frame, which is what keeps product shots and portraits usable.
Rewrite the words on a sign, a label or a poster straight from the prompt. Kontext Max is the variant to use here, since Black Forest Labs positions it as the stronger model for typography.
Extract the look of an input image, a painting style, a colour treatment or a rendering style, and apply it to a new scene while keeping the original composition readable.
With inference up to 8 times faster than FLUX 1.1 Pro upstream, Kontext is built for several rounds of small changes in a row. Each round on Aura AI costs the same flat 2 or 3 credits.
Four steps from an existing picture to a finished edit
Flat cost per edit, independent of image size or aspect ratio
| Variant | Mode | Credits per image | Best for |
|---|---|---|---|
| FLUX Kontext Pro | Image edit | 2 credits | Iterative edits, local changes, character consistency |
| FLUX Kontext Max | Image edit | 3 credits | Text editing, typography, demanding prompts |
How the four editing models in the Aura AI image editor compare
| Feature | FLUX Kontext | FLUX 2 Edit | Nano Banana Edit | GPT Image 2 Edit |
|---|---|---|---|---|
| Maker | Black Forest Labs | Black Forest Labs | OpenAI | |
| Credits per edit | 2 (Pro), 3 (Max) | 2 (Edit, Pro Edit), 3 (Max Edit) | 2 (Edit), 3 (Nano Banana 2 Edit) | 4 |
| Variants on Aura AI | 2 | 3 | 2 | 1 |
| Also text to image on Aura AI | No, edit only | ✓ Yes | ✓ Yes | ✓ Yes |
| Multi-reference editing | Single image, multi-turn | ✓ Up to 10 images (upstream) | ✓ Up to 14 images (Pro, upstream) | ✓ Yes |
| Character consistency | ✓ Core strength | ✓ Yes | ✓ Yes | ✓ Yes |
| Text rewriting inside images | ✓ Yes, best on Max | ✓ Yes | ✓ Yes | ✓ Yes |
| Plan | All plans | All plans | All plans | All plans |
Where prompt-based editing with a consistent subject pays off
Photograph one product once, then generate every colourway and every background from that single shot. The shape, label and proportions stay identical across the set.
Take a character from a story, a mascot or a portrait and place it in a different setting. Kontext keeps the face and outfit while rebuilding the environment around it.
Replace the words on a storefront, a poster or a packaging mockup without touching the surrounding design. Use Kontext Max for the cleanest lettering.
Turn a photo into a watercolour, switch daytime to golden hour, or match the look of a reference frame. The composition survives, only the treatment changes.
Common questions about FLUX Kontext on Aura AI
Prompt-based edits that keep your character and style, 2 or 3 credits per edit, no watermark. Available on every plan, worldwide.