Model Family
Create new images or make precise, context-aware edits with the FLUX.1 Kontext family. Switch between Pro and Max without leaving the page.
Pro is the default for everyday work. Choose Max when fidelity and prompt adherence matter most.
FLUX.1 Kontext Pro uses 1 credit per image.
Generated Image will appear here
Black Forest Labs built FLUX.1 Kontext to understand a written instruction together with the visual context of an image. It can generate from text, make a focused edit, or transform a full scene while keeping important relationships coherent.
Text promptNew image
Reference image + instructionContext-aware result
Where it fits today: FLUX.1 Kontext produces fixed 1MP images. BFL recommends FLUX.2 for higher resolution and newer controls.
BFL calls FLUX.1 Kontext an in-context image generation family. In practical terms, the model can start from words alone or read an existing image together with your instruction. That visual context helps it change the requested part while carrying the subject, style, layout, and surrounding relationships into the result.
60-second summary
The key idea
When you use Edit, Kontext reads the reference image alongside your words. You can target one detail, rebuild a larger part of the scene, or continue with another instruction. BFL designed this iterative workflow to preserve recognizable people, objects, and visual styles across edits.
Where it is strongest
BFL highlights local editing, character consistency, text replacement, and style transformation. These tasks all depend on the same behavior: change the requested concept while keeping the useful parts of the source image stable.
How to use it well
Name the subject, describe the change, and say what should stay consistent. For a complex transformation, make one meaningful change, review the image, then refine it with the next prompt. Quote the old and new words when editing typography.
The official variants
BFL positions Pro as the production balance and Max as the premium quality option. Its Dev release is editing-only, uses an open-weights non-commercial license, and targets local development or customization; this workspace currently exposes Pro and Max.
This pre-read condenses BFL's current model page, documentation, and knowledge-base guidance. Open the sources for API details, additional galleries, licensing, or the full model history.
Official examples
These official Black Forest Labs examples show how a short instruction can change one part of an image while preserving the surrounding visual context.
Original
ResultThe car changes color while its shape, street, lighting, and camera position stay coherent.
Prompt
“Change the car color to red”
Original
ResultKontext reads the whole portrait before removing the foreground object and reconstructing the face.
Prompt
“Remove the object from her face”
Original
ResultThe replacement follows the hand-drawn lettering, color palette, perspective, and surrounding scene.
Prompt
“Replace 'joy' with 'BFL'”
Original
ResultThe photographic scene becomes a pencil drawing while the people, framing, and layout remain recognizable.
Prompt
“Convert to pencil sketch”
Example images and prompts are sourced from the official BFL FLUX.1 Kontext documentation.
Core capabilities
Kontext is most useful when the result needs to change without losing the visual identity and relationships that already work.
Change an element while preserving its relationship with the subject, lighting, perspective, and surrounding scene.
Keep a recognizable person, object, or visual identity across scene changes and iterative edits.
Replace words on signs, labels, and posters while retaining the original placement and visual treatment.
How it works
The same workspace supports a new composition and a focused change to an existing image.
Use Generate for a new image or Edit when you want to transform one reference image.
Write a direct instruction, select Pro or Max, and choose the output shape and format.
Download the result or make another focused change while keeping the parts you already like.
Use cases
Start from a concrete job and borrow the prompt structure. Replace the subject and details with your own.
Move the same subject into a new environment while protecting identity, pose, and framing.
Try this prompt
“Change the background to a snowy city street while keeping the person, pose, and camera angle unchanged.”
Test colorways, backgrounds, packaging ideas, and campaign scenes from one product image.
Try this prompt
“Place this product on a warm beige studio sweep and preserve its exact shape, label, and materials.”
Revise a poster, sign, menu, or label without rebuilding the complete design.
Try this prompt
“Replace 'SUMMER SALE' with 'NEW ARRIVALS' while keeping the same font, colors, and layout.”
Try a new medium or art direction while retaining the source composition and recognizable subject.
Try this prompt
“Transform this image into a detailed pencil sketch while maintaining the original composition.”
Pro or Max
Both variants generate and edit. Start with Pro, then spend the additional credit only when the final asset needs Max.
FLUX.1 Kontext
The balanced choice for everyday generation, focused edits, and quick creative iteration.
1 credit per image
FLUX.1 Kontext
The quality-first choice for demanding prompts, typography, and polished final assets.
2 credits per image
Prompt guide
Precise edit prompts reduce accidental changes. A useful prompt names the target, describes the action, and identifies the visual details that should remain stable.
Focused instructions are easier to control. Build complex edits through a short sequence.
Use direct phrases such as 'the red car' or 'the person in the foreground' instead of vague pronouns.
Call out the face, pose, product shape, composition, lighting, or background details that must remain stable.
For typography edits, put the original and replacement words in quotation marks.
Review each result, then use the next instruction to refine one additional part of the image.
FAQ
The practical details to know before you spend a credit or choose another FLUX family.
FLUX.1 Kontext is Black Forest Labs' image model family for both text-to-image generation and context-aware image editing. It reads a written instruction together with an optional image so it can create a new result while preserving important visual relationships.
Start with Pro for everyday generation, fast iteration, and most edits. Choose Max when stronger prompt adherence and final-image fidelity justify the additional credit.
Yes. Select Generate to create from text, or Edit to upload one reference image and describe the change you want.
FLUX.1 Kontext Pro uses 1 credit per generation. FLUX.1 Kontext Max uses 2 credits per generation. The selected cost is shown before you submit.
This workspace accepts one reference image in Edit mode and returns a fixed 1MP result. You can select the available aspect ratio and export as JPG or PNG.
Choose FLUX.2 when you need the current FLUX generation, output up to 4MP, newer controls, or a workflow that will later expand to multiple references. FLUX.1 Kontext remains useful for focused, single-reference generation and editing.
Commercial use depends on the selected model and the Kontext AI terms that apply to your account and project. Review those terms before publishing or selling an output.
Use Pro for your first pass, switch to Max for a quality-first final result, and refine one deliberate change at a time.
