FAST PORTRAIT CUTOUT
Upload a photo, review the outline and save a transparent PNG in one focused workflow.
Drop picture here
or click to browse · Ctrl+V to paste
JPEG · PNG · WebP
Upload a photo, review the outline and save a transparent PNG in one focused workflow.
Check flyaways, curls, glasses, collars and shoulder contours before export.
Quick local inference with no image upload queue. Actual time varies by device, image size and first-time model loading.
Drop picture here
or click to browse · Ctrl+V to paste
SoloBG focuses on human portraits, where a convincing background removal depends on small boundaries: loose hair, facial contours, glasses, shoulders and clothing. Start with a sharp source, inspect the result at zoom and export a transparent PNG. Lighting and background contrast still affect the estimate.
Loose strands can blend into nearby colours. A sharp, well-lit photo gives the model a clearer boundary to follow.
Check eyewear, ears and cheek contours at zoom before placing the portrait over a new background.
Hoods, lapels and patterned clothing can merge with the scene when their colours or shadows are similar.
A small, blurry or compressed photo can soften the cutout edge. Enlarging the export cannot restore missing detail.
The transparent PNG keeps the source pixel dimensions where browser memory allows, with no watermark added by SoloBG.
The cutout is processed locally, not uploaded to a SoloBG image-processing server. Model assets still load from their hosts.
Prepare a portrait cutout for a website bio, staff directory, portfolio or community profile.
Place a person over video thumbnails, podcast artwork, course covers and social graphics.
Export a still image as a PNG for compositing in a video editor or presentation; SoloBG processes images, not whole videos.
Use a person wearing a product in a cleaner composite for a listing or campaign design.
Add a new background, crop, shadow or colour treatment in the image editor you already use.
No. The AI model is downloaded to your browser once and cached locally. All portrait matting runs in a background thread on your own device. Your images never leave your machine and we never see them.
The MODNet portrait matting model is approximately 12 MB and is fetched directly from Hugging Face on first visit. The browser may cache the model assets, so repeat visits can require fewer downloads depending on the browser cache and storage settings.
SoloBG accepts JPEG, PNG, WebP, and most browser-readable image formats. For best results use the highest-quality version of the image available. Heavy JPEG compression can affect edge sharpness.
SoloBG uses a portrait matting model optimised for single-subject foreground isolation. Group photos with multiple people will include all visible people in the foreground. For best results, crop to a single subject before uploading.
The downloaded PNG uses the input image’s pixel dimensions rather than the on-screen preview dimensions, and SoloBG does not add a watermark. A 4K input can produce a 4K canvas when browser and device memory allow; output dimensions do not recreate detail missing from the source.
Yes. The tool is fully responsive and the model runs on any modern browser including Chrome and Safari on iOS and Android. Processing will be slower on mobile devices without GPU acceleration.
SoloBG is a processing tool — you retain all rights to your images. The underlying model is licensed under Apache 2.0. There are no watermarks and no output restrictions from our side.