Feature Guide
Object Editor
Edit just one thing. Generate or upload an image, let the AI find its editable parts, then click any of them to replace, restyle, remove, or re-text — everything else stays untouched.
At a glance
Everything you need to know
Three ways to start
Generate a fresh canvas from a prompt, upload a PNG/JPG (up to 8 MB), or pull in something you already made from Your Gallery.
Click-to-edit areas
One tap maps the image into clickable regions — text, subjects, objects, and the background. Missed a spot? Draw a box or paint one with the brush.
Four edit kinds
Replace (free-text), Remove (generative fill), Restyle (8 surface presets), and Edit Text — swap the words while the font, size, and lighting stay put.
Character swap
Replacing the main character? One toggle updates every identity-dependent detail across the whole image — antlers become ears, the latte art follows suit.
Before/after compare
After every edit the canvas opens a drag slider comparing the new version with the one it came from — see exactly what changed before you continue.
History that branches
Every version lives on a forking tree. Jump back to any version and branch a new direction — nothing is ever overwritten or lost.
The core loop
Click. Describe. Done.
Most edits follow the same three beats: pick the part, describe the change, run it. The action bar appears next to your selection with everything you need.
- Replace — the box is prefilled with the AI’s description of what’s currently there; edit it into what you want instead (“cream armchair” → “blue velvet armchair”).
- Remove — erases the selection and fills the gap with matching surroundings.
- Restyle — keeps the shape but changes the surface: Photorealistic, Watercolour, Flat Vector, Neon Glow, and more.
- Edit Text — the AI reads the existing wording; you type the replacement. Typography, colour, and lighting are preserved.
Floating action bar
Replace · Remove · Restyle · Edit Text
Precision
Three ways to select
Overlapping boxes are handled for you — clicking a headline that sits inside the background selects the headline, because the smallest region under your click always wins.
- AI areas — “Find editable areas” returns up to a dozen labelled regions you can click or pick from the list.
- Box tool — drag a rectangle over anything the AI missed.
- Brush tool — paint the exact shape (with eraser and size control), or use “Snap mask to edges” to tighten an AI box around the object’s real outline.
Box & brush
Draw or paint custom areas
Character swaps
Change the star, keep the scene
Swapping a reindeer for a bear isn’t a local patch — the antlers, the species patterns on the sweater, and the reindeer face in the latte foam all belong to the old character. When the selected region is the main subject, the Character swap toggle (on by default) updates every identity-dependent detail across the whole image, while the composition, background, lighting, and text stay untouched.
- · Species features become the new character’s features
- · Clothing and props keep their style, re-proportioned for the newcomer
- · Repeated depictions (latte art, motifs, ornaments) follow the swap
- · Existing text is left alone — re-word it with Edit Text afterwards
Character swap
Reindeer → Bear, everywhere
Confidence
Variants & compare
Ask for 1–3 variants per edit. A single result applies straight to the canvas; multiple results show a candidate grid where you can drag a before/after slider on each before choosing Set as canvas.
The compare slider also opens automatically after every applied edit, and the toolbar Compare toggle brings it back any time you want to see what changed since the previous version.
Before / after
Drag to reveal the change
Walkthrough
How to use the Object Editor
Bring in an image
Generate from a prompt, upload a file, or pick a recent generation from Your Gallery.
Find editable areas
One credit maps the image into clickable regions. Or skip this and draw your own with the Box / Brush tools.
Select a region
Click it on the image or in the areas list. The floating action bar opens beside it.
Choose the edit
Replace (prefilled with the current description), Remove, Restyle, or Edit Text. Attach reference chips to steer the look.
Run it
1–3 variants; single results apply instantly, multiple results land in the candidate grid.
Compare & continue
The before/after slider opens automatically. Star the version to Your Gallery or download it.
Fork anywhere
Click any version in the History timeline to make it the starting point for your next edit — old branches stay clickable (dashed).
Tips
Good to know
Credits. Finding editable areas costs 1 credit; each edit variant costs the same as a normal generation with your chosen model — asking for 3 variants is 3 generations.
Text edits keep the typography. The original text stays visible to the AI so the replacement matches its font, size, colour, and lighting — only the words change.
Bring your own image. Uploads work on anything — photos, AI art from other tools, scans — up to 8 MB, and the aspect ratio is detected automatically.
Edges matter. If an edit came out sloppy, try “Snap mask to edges” first — a tight mask around the object beats a loose box every time.
Star to save. Versions only appear in Your Gallery when you star them; everything else stays in the editor’s history for the session.
Clicking empty space deselects. The smallest region under your click wins, so dense images with overlapping areas still behave predictably.
Ready to create?
Open Object Editor in the Responsive Templates designer and build your design in minutes.