Skip to content

Pick your style: Denkbaar

Book a conversation

Articles

History

FLUX.1 Kontext edits images using text and image input

·

Black Forest Labs introduced a model family for creating and selectively editing images.

A designer compares two images of the same tool.
Record what changes and what must remain unchanged. Illustration: Denkbaar.

On 29 May 2025, Black Forest Labs introduced FLUX.1 Kontext. The model family combined text and images as input for creating and editing images. The maker emphasised targeted changes and keeping subjects recognisable. The announcement explains the approach. An image and a brief card lead to an edited image.

For image editing, the assignment itself is a useful starting point here. You could start from an existing image and describe what needed to change. This made it useful to record both the requested change and what must remain unchanged. A magnifying glass lies beside three drawings of the same spanner.

A small or medium-sized business could try this with an original illustration for a manual. For example, request a different background while keeping the depicted tool’s shape the same. Use material for which you have the necessary rights. Place the original and edited images side by side. Check components, proportions and small details. Ask someone to assess whether the instruction still explains exactly the same thing. Try a second edit too, to see whether the subject stays recognisable. This assesses image editing through a concrete assignment and a useful sequence, without assuming the promised consistency beforehand.

Sources

  1. FLUX.1 Kontext (in-context image editing)

Topics: image/video