Skip to main content

Mask

Create Mask

Create precise object masks using simple text prompts—no manual selection required.

Overview

The Create Mask tool allows you to automatically isolate any object in an image by describing it in plain text. Fluidfit’s AI detects the subject and generates accurate masks in seconds, helping you work faster with full control over each element.

You can also invert masks or combine multiple masks into a single mask—all without manual drawing.

What Can You Do With Mask

  • Object-level segmentation for people, products, apparel, faces, and backgrounds

  • Background removal and replacement with clean, accurate edges

  • Targeted effect application (color, blur, lighting, texture, style) on specific regions

  • Multi-object control with options to invert or combine masks

  • Advanced compositing support for cutouts, overlays, and scene assembly

How It Works

1. Drag & Drop Media Node: Add the Media node to the canvas and upload your source image.

2. Add Create Mask Tool: Drag the Create Mask tool and connect it to the Media node.

3. Enter Prompt: In the Create Mask tool, describe what you want to isolate (e.g. face, dress, person, background).

4. Invert & Compositing Mask (Optional)

  • Enable Invert Mask to select everything except the defined subject.

  • If multiple objects are detected, use Composite Masks to merge them into a single mask for a single output.

5. Generate Output: Click Generate to get AI-powered, precise masks with full control.

Note: If there are more than 1 objects identified, you will get all those masks separately to provide you the better control on each object and use Compositing Mask feature to merge them into a single mask.

Prompt Examples

  1. Use clear, simple object names for best results: person, face, dress/fabric , shirt, handbag, shoes, background, car, bottle

  2. Avoid long sentences—keep prompts short and clear

  3. If results are not accurate, try: "More specific terms (e.g. “white shirt” instead of “shirt”)"

    Tip: The more specific the prompt, the better the mask accuracy.

This feature is also available via API. Refer to the API documentation here for implementation details.


Try Fluidfit Today

Did this answer your question?