Studio Image Editing

Overview

Studio Image Editing is the core workflow for turning one source image into publishable visual options. Start from an uploaded image, a generated result, a product photo, a screenshot, or an imported PSD. Then select the image on the canvas and choose the editing action that matches the job.

The important difference is that Studio does not treat every edit as the same prompt box. It separates conversational edits, marked-up local edits, editable reconstruction, cleanup, background replacement, variation generation, sizing, and export preparation so each step has the right controls.

Where Image Editing Starts

Entry point	Use it for
New Editor quick start	Upload an image and jump directly into AI Chat, Make Editable, Edit Element, Remove BG, Eraser, Variations, Upscale, Expand, or Vector.
Generated result	Continue from an image created in a Studio session and refine it without losing the prompt history.
Canvas selection	Select any supported image on the canvas and open the image action panel.
PSD import	Bring a layered or flat Photoshop asset into Studio, then edit, reconstruct, or export from the canvas.
Library asset	Reuse an uploaded or generated image from Library and continue editing it in a new context.

Pick The Right Editing Path

Goal	Best path	Result behavior
Change a selected image with a written instruction	Chat Edit	Creates an edited version near the source for comparison.
Point to exact areas and explain what should happen	Visual Annotation Edit	Creates a new result using your marks as edit instructions.
Turn a flat image into editable objects	Make Editable or Magic Layers	Creates an editable frame beside the source.
Replace or remove visible text in an image	Edit Design Text	Detects text blocks, then creates a regenerated image with edited text.
Isolate a product or subject	Remove BG	Updates the selected image with a transparent-background result.
Replace a product background	Set Background	Uses local background layers when possible, or AI image editing for generated/blurred backgrounds.
Remove unwanted objects	Eraser	Uses a painted mask and updates the selected image.
Explore more versions of the same subject	Generate Variations, Remix, or Multi-angle	Creates new options beside the source.
Prepare for a larger placement	Upscale or Expand	Creates a larger or reframed version beside the source.
Convert suitable artwork into vector output	Vector	Creates a vectorized version beside the source.
Recover the likely source prompt	Extract Prompt	Writes an editable prompt you can copy or reuse.

Directed Edits

Use Chat Edit when you know the change but do not need manual object-level editing. The editor sends the selected image, the prompt, optional mask data, optional reference images, and generation settings together. This is best for changing mood, lighting, styling, localized details, or visual composition while preserving the source.

Use Visual Annotation Edit when words alone are ambiguous. Mark the image with brush strokes, circles, rectangles, or delete marks, add notes, choose a model, and generate. Studio uses the annotated reference as an instruction layer and removes the marks from the final output.

Good edit prompts include two parts:

Editable Reconstruction

Use Make Editable when the next step is object-level editing. Before conversion, you can crop the source area or tag important regions as image areas, then Studio creates an editable design frame beside the original image.

Use Magic Layers / Edit Element when a flat image needs to be separated into structured layers quickly. This is useful for screenshots, UI images, posters, generated graphics, product cards, and image assets where you want to move, restyle, or remove individual pieces after extraction.

Choose editable reconstruction when you need to select and move things. Choose Chat Edit when you only need a new rendered bitmap.

Background And Cleanup

Studio includes separate tools for the most common production edits:

Tool	Best for
Remove BG	Isolating a product, person, object, or icon from its background.
Set Background	Applying a color, gradient, uploaded background, searched image background, curated texture, AI background, or blurred contextual background.
Eraser	Removing a small object or distracting area with a painted mask.
Remove Watermark	Cleaning assets you own or are allowed to modify.
Adjust Colors	Local brightness, contrast, saturation, and opacity finishing.
Crop	Non-AI framing cleanup before or after generation.

For ecommerce or product work, a stable workflow is: remove background, set a neutral or brand-aligned background, adjust crop, generate variations, then upscale the selected version.

Text, Variations, And Output Size

Edit Design Text detects text blocks in an image, lets you rewrite or remove them, and regenerates a version that preserves the original composition and typography style as closely as possible. Use it for posters, covers, banners, social graphics, and AI images with incorrect text.

Generate Variations creates 1 to 4 alternatives with control over strength, aspect ratio, resolution, and model. Use low strength for small refinements and stronger settings when exploring new directions.

Remix restyles the source image while keeping the subject recognizable. Multi-angle generates a new camera viewpoint with rotation, tilt, and zoom controls. Expand reframes an image for social and marketplace formats, while Upscale prepares a sharper or higher-resolution version.

Review Checklist

• Check product details, faces, hands, logos, prices, and visible text before publishing.
• Use masks or annotations when only a small region should change.
• Keep one primary change per edit for better control.
• Preserve brand and product identity explicitly in the prompt.
• Use Remove Watermark only on content you own or have permission to edit.
• Keep the original image on the canvas until the final version is approved.

Related

• Image Editing Actions
• Image Editing Workflows
• Canvas AI Actions
• Prompt And References