Higgsfield has unified the entire creative stack!
Photo editing, camera motion, VFX, and avatar speech — all in one pipeline with Flux.1 Kontext
This is fire.
Here's how to use Kontext: 🧵
1) Getting Started
→ Start by heading to @higgsfield & select "Create"
https://higgsfield.ai/create/image
2) Kontext prompting guidelines
→Say what not to touch.
If something should stay the same, call it out.
→Be specific.
Say exactly what you want. Skip the vague stuff.
→Use clear action words.
Go with “change” or “replace” instead of “transform” — it’s easier to follow.
→Take it step-by-step.
Break big changes into smaller edits. One thing at a time.
3) Tips for best common use cases
→Edit text:
“Replace ‘[old text]’ with ‘[new text]’, but keep the font and style as is.”
→Swap an object:
“Change the [object] to a [new thing], but leave [what stays the same] alone.”
→Swap backgrounds:
“Replace the background with [new background], but keep the subject in the same position.”
→Switch styles:
“Make it look like [style], but keep the [pose / layout / character] the same.”
4) Typical user mistakes
→Not getting character consistency
Bad Prompt: “Change the man into a king”
Good Prompt: “Swap the man's clothes to a king's outfit, but do not change the man's face”
→Style change isn't working
Bad Prompt: “Make it a watercolor style”
Good Prompt: “Change this image into a watercolor painting with with hand strokes, strong contrasts, and vibrant colors”
5) That's it for now! I have partnered with Higgsfield to experiment with Kontext. You can try it yourself here:
https://higgsfield.ai/create/image
6) if you enjoyed this thread be sure to like, bookmark, and repost!
https://x.com/Mr_AllenT/status/1933322810900033922
