flux-kontext

To perform local image edits without losing original identity, use Flux 1 Kontext Pro for single-reference image-to-image processing. It applies prompt-driven photo retouching to update specific objects while preserving source identity, pose, and framing.

The best way to swap objects or text in a single photo using Flux is single-reference editing with the Flux 1 Kontext Pro model. It executes brand and text swaps while maintaining the original image composition through preservation-first prompting patterns.

Yes, you can use Flux for prompt-driven photo retouching on just one image. The Flux 1 Kontext Pro model processes single-reference local edits, allowing you to add, swap, or modify specific elements without altering the underlying image structure.

Yes, Flux image-to-image editing supports optional aspect ratio and seed controls. When invoking the RunComfy CLI with the Flux 1 Kontext Pro model, you can include these parameters in the JSON input schema to generate seeded variants of your edited images.

You should use Flux Kontext Pro for precise, single-image local edits that preserve source identity. If your task requires multi-image edits, embedded text changes, or text-to-image generation, you should switch to related RunComfy edit models for better results.

Flux image edits drift away from the original pose and framing due to naive prompting. Using preservation-first and single-instruction edit phrasing with the Flux 1 Kontext Pro model improves consistency and keeps the source identity intact during local edits.