AI Image Studio — Object Remover & Background Eraser
Paint a mask to remove objects, cut out backgrounds, upscale 2x or 4x and expand images in your browser. An optional cloud key uploads the image.
Still loading. If the tool doesn’t appear, reload this page.
About this tool
AI Image Editor — Object Remover, Inpainting & Background Eraser
Paint a mask to remove objects, cut out backgrounds, upscale 2x or 4x and expand images in your browser. Optional cloud keys upload the image to that provider.
This editor combines mask painting, object removal, background cutout, image expansion, 2x or 4x upscaling and click-based Smart Mask in one browser canvas. The built-in tools process pixels on your device with Canvas 2D, so no image is uploaded. Background removal uses the @imgly/background-removal model, which downloads from staticimgly.com the first time you use it: about 96 MB at High Quality, the default, or about 54 MB at Standard. If you save your own OpenAI, Stability AI, ClipDrop or Hugging Face key, choosing that provider uploads the image and mask to it (api.openai.com, api.stability.ai, clipdrop-api.co or router.huggingface.co). The layout follows the open-source IOPaint project, which does not run here.
How to use AI Image Studio — Object Remover & Background Eraser
- 1
Drop a PNG, JPG or WebP photo into the workspace or click Browse Image to open one up to 4096 by 4096 px.
- 2
Paint over unwanted objects, watermarks, or blemishes using the Mask Brush (B), Rectangle (M), or Lasso (L) selection tools.
- 3
Choose Remove Object, Remove BG, Expand, Upscale 2x or 4x, or Smart Mask in the right-hand AI panel and click the primary action button.
- 4
Inspect the result with the Compare split slider or hold Tab to peek at the original photo, and refine mask edges using Expand +10 or Shrink -10.
- 5
Switch to the Export tab, pick PNG, JPEG, or WebP with optional quality compression, and download the edited image to your device.
Why use AI Image Studio — Object Remover & Background Eraser
Built-in tools process pixels on your device
Object removal, mask expand and shrink, reflected expansion, Smart Mask and PNG, JPEG or WebP export run in your browser without uploading the image.
Background cutout with a fallback
Remove Background runs @imgly/background-removal in the browser after downloading its model from staticimgly.com once. If the download fails, a border-colour matting pass runs instead.
Precision masking and undo history
Combine soft brush strokes, an eraser, rectangle and lasso selections, mask invert, expand or shrink by 10 pixels, and 30 undo steps.
Optional cloud providers with your own key
Add an OpenAI, Stability AI, ClipDrop or Hugging Face key to run prompt-driven fills. The key is stored unencrypted in this browser and the image goes to that provider when you run a task.
Common use cases
- Remove unwanted bystanders, power lines, logos, timestamps, or blemishes from travel photos and product shots with multi-pass onion-peel inpainting.
- Cut out foreground subjects with crisp transparent alpha channels for e-commerce listings, thumbnails, stickers, and presentations.
- Expand tight portrait or landscape framing along any edge using reflected border outpainting and interactive canvas drag handles.
- Upscale low-resolution graphics, UI captures, or illustrations 2x or 4x with multi-step resampling and 3x3 unsharp contrast sharpening.
Frequently asked questions
Does it work without an API key?
Yes. Object removal, boundary fill, upscaling, expansion, Smart Mask, mask editing and export run in your browser with no key. Remove Background also needs a first-use download from staticimgly.com: about 96 MB at High Quality, the default, or about 54 MB at Standard.
Is my image uploaded anywhere?
Not by the built-in tools. It is uploaded only when you save a provider key and run a task on that provider: api.openai.com, api.stability.ai, clipdrop-api.co or router.huggingface.co. In Auto mode a saved key is used for Inpaint and Object Replacement.
How does background removal work?
It runs the @imgly/background-removal model in your browser to produce a transparent PNG. If the model cannot be downloaded, a simpler border-colour matting pass is used, which suits plain backgrounds.
How does object removal fill the masked area?
A small in-browser fill copies colour inward from the edge of the mask layer by layer and adds light grain. It suits small objects, text and watermarks on simple backgrounds. It is not a neural model and it ignores the prompt.
Where are API keys stored?
In this browser's localStorage under altft_ai_image_editor_keys, unencrypted, and sent only to the provider you choose. Use Clear in the Model tab to delete them.
Is this IOPaint?
No. It follows the open-source IOPaint workflow by Sanster, but IOPaint and its LaMa model do not run in this tool.
What image formats and sizes are supported?
PNG, JPG, JPEG and WebP up to 25 MB and 4096 by 4096 pixels. Export as PNG with alpha, JPEG flattened on white, or WebP.
Which keyboard shortcuts are available?
B for Brush, E for Eraser, H or Space for Pan, M for Rectangle, L for Lasso, [ and ] to resize the brush, + - 0 to zoom, Tab to peek at the original, Ctrl or Cmd+Z to undo.