Answer

Short answer:

Regenerating creates a completely new image or interpretation based on your inputs, effectively starting the rendering process from scratch. Editing modifies specific parts of an existing image, such as changing a material or removing an object, without altering the overall composition or aspect ratio.

1/ Global creation vs. local refinement

The distinction between these two actions is about where you are in the design workflow.

Regenerating is an exploration tool. It is used when you are still defining the look, lighting, or structure of a scene. Even if you use a base image (like a sketch or 3D screenshot), regenerating tells the AI to interpret that input and synthesize a fresh set of pixels. This often results in global changes to lighting, texture, and atmosphere.

Editing is a precision tool. It assumes you are happy with 90% of the image and need to fix the remaining 10%. It uses techniques like "Select & Modify" to seamlessly blend new elements into the existing pixels, preserving the perspective and context of the original render.

> “I like the mood, but the client hates that specific chair.” , This is an editing task.

> “I don’t like the mood at all, let’s try a sunset.” , This is a regenerating task.

2/ How it works

Regenerating

When you regenerate (using tools like Text-to-Render, Sketch-to-Render, or 3D-to-Render), the AI builds the image from the ground up.

1. Input: You provide a prompt and optional visual guides (sketch, elevation, or 3D base).

2. Process: The AI generates a new visual interpretation. It determines the camera angle, lighting, and material application simultaneously.

3. Outcome: A fresh image. If you run the same prompt twice, you will get two slightly different variations.

Editing

When you edit (using the Edit tools), the AI is constrained by the existing image.

1. Input: You select an existing render and use a tool like a brush or lasso to mask a specific area.

2. Process: The AI analyzes the surrounding pixels (context) and only generates new content inside the masked area. It tries to match the lighting and perspective of the unmasked parts.

3. Outcome: The original image remains identical, except for the specific area you changed.

3/ Capabilities

Regeneration supports:

• Drastic style changes: Switch from "minimalist concrete" to "warm timber" in one go.

• Compositional exploration: See how a sketch looks as a photograph.

• Aspect ratio selection: Choose specific ratios (1:1, 2:3, 16:9) for text-to-render workflows.

• Base interpretation: Turn a loose line drawing into a fully realized building.

Editing supports:

• Object removal: Erase cars, people, or artifacts that distract from the design.

• Material swapping: Change a rug pattern or wall color while keeping the lighting consistent.

• Correction: Fix AI artifacts (like a distorted table leg) without losing the rest of the image.

4/ When to use which

Use Regenerating when:

• You are in the concept phase and want to see 5 different atmospheric options.

• The current image is "not quite right" globally and needs a fresh start.

• You are moving from a low-detail input (like a massing model) to a high-detail visualization.

Use Editing when:

• The render is perfect, but there is a floating artifact in the sky.

• You need to show a client two variations of furniture in the same room.

• You want to increase the resolution of a final selected image.

• You need to extend the edges of an image (if supported by the specific tool).

5/ Limitations and notes

• Resolution: Default generations are typically around 1MP and editing tools generally use the same image sizes.

• Credit Consumption: Both actions consume credits, but they serve different purposes. Using "Regenerate" to fix a small error is often inefficient; use "Edit" instead to save time and maintain consistency.