Skip to Content

Inpainting Images

Inpainting is a specialized image editing technique that allows you to make targeted changes to specific areas of an image while preserving the rest. Bearly’s inpainting feature gives you pixel-level control for sophisticated image edits.

What is Inpainting?

Inpainting is the process of reconstructing or modifying selected parts of an image. Unlike full image generation or editing, inpainting focuses only on the areas you specifically mark, leaving the rest of the image untouched. This technique is perfect for:

  • Removing unwanted objects
  • Filling in missing details
  • Changing specific elements without affecting the overall composition
  • Making creative, targeted modifications

Using the Inpainting Tool

Click the brush on the image you want to edit.

Click the brush icon on any image in your chat that you want to modify.

Select image

Paint the target area

Use your cursor to draw over the specific areas you want to modify. The bright green highlight shows your selection mask.

  • Use broad strokes for larger areas
  • Be precise for detailed edits
  • You can use undo/redo buttons (or Cmd+Z/Ctrl+Z) to refine your selection

Paint target area

Write your instruction prompt

Enter a clear, detailed description of what you want to change in the highlighted area.

Examples:

Replace with a blue sky with fluffy clouds
Change to a marble countertop
Add a sleeping cat

Write prompt

Submit & Review the result

Hit submit!

The edited image will appear in your chat with your changes applied to the targeted area while preserving the rest of the image.

Review result

Advanced Inpainting Techniques

Precise Masking

For the best results, be deliberate about your mask area:

  • Complete coverage: Ensure you fully cover the area you want to change
  • Clean edges: Try to create clean boundaries at the edges of your selection
  • Context awareness: Include a small amount of surrounding area if you want the edit to blend naturally

Effective Prompting

Your prompt instructions significantly impact the quality of inpainting:

  • Be specific: “Replace with a red rose” works better than “add a flower”
  • Include style guidance: Mention “photorealistic,” “cartoon style,” or “oil painting” to match the original image
  • Reference surrounding elements: “Add a chair that matches the table” helps create coherent edits

Multi-Stage Editing

Complex edits often work better as a series of smaller inpainting operations:

  1. Make one targeted change and submit
  2. On the resulting image, make another inpainting edit
  3. Continue refining until you achieve the desired result

Technical Implementation

Bearly’s inpainting feature uses OpenAI’s GPT-Image model which excels at understanding both visual context and text instructions. When you submit an inpainting edit:

  1. Your brushstrokes create a binary mask identifying the target area
  2. This mask and the original image are processed alongside your text prompt
  3. Only the masked portion gets regenerated based on your instructions
  4. The result is seamlessly integrated with the unmasked portions of the original image

Practical Applications

  • Photo correction: Remove blemishes, unwanted objects, or distractions
  • Creative editing: Change colors, textures, or add new elements
  • Design iterations: Test different options for specific components
  • Composite creation: Blend elements from different visual styles
  • Background replacement: Change settings while keeping subjects intact
Last updated on