Remove Object

Catalog

• Display name: Remove Object
• Categories: Utility
• Status: Ready
• Route label: GPT Image 2 consistency-sensitive edit
• Tone: red
• Short description: Remove one object and naturally rebuild what was behind it.
• Similar skills: clean-room, restyle-room

What this skill does

• Lets the image model identify one target from language and infer the hidden background in context.
• Keeps crop, perspective, lighting, texture, people, and unrelated objects stable.

How to use

Provide image and identify the target by type, position, color, and relation to nearby objects.

Remove one distraction

Remove only the red trash bin at lower right and reconstruct the pavement and wall behind it.

Workflow

1. Bind one removal target unambiguously and list all protected neighboring content.
2. Describe the expected continuation behind it from visible scene cues, without inventing a new focal object.
3. Run gpt-image v1.0 with quality: "medium", imageSize: "2K", the supplied references, the requested or source-preserving ratio, batchCount: 1, and one result.
4. Ask for one contextual edit; forbid crop, relighting, global cleanup, extra people, or unrelated changes.
5. Inspect target absence, boundary residue, repeated texture, geometry, shadows, and protected landmarks; retry the failed region only.

User-facing output

• Media type: One repaired image
• Default quantity: 1
• Content per image: Source scene with one target removed
• Default layout: Source composition
• Model policy: GPT Image 2 Medium/2K for consistency-sensitive final image editing
• Downstream use: Photo cleanup and commercial retouching