Background Remover

BETA

Remove a portrait background, tune the cutout, and export a full-size image on your device.

Quick answer and specifications

The core workflow runs in this browser. Check the supported formats, published limits, and any on-demand model download before you start.

Inputs
JPG, PNG, & WebP
Outputs
PNG, WebP, & JPG
Processing
On this device, in the browser
Per-file limit
20 MB
Files per run
1
On-demand model download
24 MB
Last verified

Your image and generated mask stay on this device. The model runs in a browser worker.

This model is optimized for people and portrait photos. Products, animals, and complex scenes may not cut out cleanly.

Drop or paste a portrait image

JPG, PNG, or WebP, up to 20 MB and 50 megapixels.

Select files

How to remove a portrait background locally

Aiviko's Portrait Background Remover isolates people and portraits from JPG, PNG, or WebP images up to 20 MB and 50 megapixels directly in your browser without uploading the photo. It runs a local MODNet portrait matting model (about 25 MB downloaded on first use and cached) inside a browser worker with WebGPU acceleration and WASM fallback, then exports full-resolution PNG, WebP, or JPG files.

  1. 1. Drop or paste a portrait image (up to 20 MB)

    Choose a well-lit JPG, PNG, or WebP portrait photo up to 20 MB and 50 megapixels.

  2. 2. Run the local 25 MB MODNet portrait model

    On first run, the browser downloads and caches the ~25 MB ONNX model, then computes the alpha matte in a worker via WebGPU or WASM.

  3. 3. Refine edges, pick a background, and export

    Inspect the before/after slider, fine-tune edge shift, feather, and opacity, choose transparent, solid color, or blurred background, and download full-size PNG, WebP, or JPG.

Supported portrait cutout and compositing features

  • Local AI portrait background removal for JPG, PNG, and WebP photos up to 20 MB and 50 megapixels (within a 768 MB decoded memory budget)
  • Interactive before/after comparison slider plus conservative edge refinement controls (expand/contract shift, feathering, and alpha opacity strength)
  • Flexible background modes: transparent cutout, solid white, solid black, custom hex color picker, or adjustable Gaussian blur of the original background
  • Full-resolution export in PNG or WebP (preserving transparency) or JPG (composited onto an opaque background)

Portrait model scope and format limits

  • The bundled MODNet model is trained specifically for human portraits; products, pets, cars, or complex multi-object scenes may not separate cleanly.
  • JPG does not support alpha transparency—export as PNG or WebP when you need a transparent cutout.
  • Fine flyaway hair, semi-transparent veils, motion blur, or subjects matching the background color may require conservative edge feathering.

Frequently asked questions about Background Remover

Are my portrait photos uploaded to a cloud server?

No. Your source image, generated alpha mask, edge adjustments, and final full-size export stay inside your browser. Only the ~25 MB MODNet model and ONNX runtime are downloaded on first use.

Can I remove backgrounds from product photos or iPhone HEIC files?

This model is optimized for people and portraits rather than general products or animals. If your portrait is in iPhone HEIC format, convert it to PNG or JPG in Aiviko's File Converter first.

Which export format keeps a transparent background?

Download PNG or WebP when Transparent mode is selected. JPG files do not support transparency and will export on a solid white (or your selected) background.

Before you use this tool

Product information reviewed by Aiviko Team

Where processing happens

The selected input is processed in your browser and is not uploaded to an Aiviko processing server.

What may use the network

The page may download code, fonts, runtime files or machine-learning models. Analytics and advertising load only under the choices and conditions described in the privacy and cookie policies.

Results and limits

Speed, accuracy and maximum practical file size depend on the input, browser and device. Review the guide before using the output for an important decision.

Read the full guide and troubleshooting notes →