⌘V from clipboard if you've just copied from Capture One).If you already know what you need, jump straight to the tool. No signup, no upload to a server, no watermark on the output.
Try the Background Remover →You shoot 40 corporate headshots on Tuesday. The marketing director needs three teaser images for LinkedIn that afternoon — knocked-out subject on a flat brand color so they sit cleanly inside the carousel template. Cutting all three by hand in Photoshop would take 15 minutes. Doing them in the browser, downloading PNGs, and dropping them into a Figma frame the marketing team can grab takes under two minutes. The retouching deadline for the full delivery doesn't move.
You shot 60 SKUs for a small DTC apparel brand. The merchandising team wants to see the product on a pure-white PDP layout before you finalize retouching. Run the JPEG previews through the cutout tool, drop them on a #FFFFFF background, screenshot for the client. They sign off on framing and prop choices on Wednesday instead of waiting until Friday's retouched delivery to flag a hanger or shadow they don't like. Your reshoot risk drops to near zero.
For your own site, a clean cutout of a flagship product or a portrait subject sits well on a hero section. Pair it with a brand-aligned background color from the palette tool and you have a portfolio teaser without renting a studio cyc wall. Same trick works for blog posts and Substack newsletter art when you're writing up a shoot retrospective.
Real estate photographers often need a quick subject-isolated shot of a kitchen island, a furniture piece, or a staged room for the listing agent's social posts. Cut out the focal element, place it on a brand-colored card, and the agent has a scroll-stopping carousel for Instagram Reels covers without you delivering 40 more retouched frames.
You're shooting a product launch. PR wants three hero shots of the founder holding the product, isolated, for the embargo-day press kit. Cull, export JPEGs, cutout, pass to PR. Done in the 90 minutes between the event ending and the press embargo lifting. The full edit gets delivered the next morning.
| Use case | Verdict | Why |
|---|---|---|
| Social teasers, IG carousels | Ship it | Compression on social platforms hides minor edge artifacts. |
| PDP comps for client sign-off | Ship it | Speed of feedback beats pixel-perfect masking at the comp stage. |
| E-commerce final hero shot | Touch up after | Run through, then refine the mask in Photoshop or Capture One. |
| Print campaign at 300dpi | Touch up after | At A3 print scale the eye finds 1-pixel artifacts. Always do a final pass. |
| Hair on busy backgrounds | Manual mask | Translucent strands need either a paid tool (Topaz Mask AI, Photoroom) or a hand-refined mask. |
| Glass, smoke, water, motion blur | Manual mask | Anything genuinely translucent is a hand-mask job in any tool, free or paid. |
Three reasons it's worth a browser tab instead of a desktop subscription or an online upload service:
The cutout is rarely the final destination. Here's how working photographers wire it into a typical post-production pipeline:
The model is good at clear subject-background separation. It's less good at three things photographers care about: wispy hair on backgrounds tonally close to it, fine jewelry chains, and dark rim lighting that softly blends a subject's edge into a dark backdrop. None of these are unique failure modes — they're the same edge cases that defeat every automatic cutout product on the market. The honest workflow is: do the auto cutout, identify which 10% of frames have edge issues, hand-refine only those in Photoshop, and ship the other 90% as-is. That keeps your effective rate at $300+/hour instead of $80/hour.
The output is a PNG at the input's pixel dimensions, capped at the browser's canvas limit (typically 16384px on the long edge — enough for any DSLR or mirrorless body shooting at native resolution). EXIF metadata is not preserved on the PNG output — that's an inherent limit of the PNG-from-canvas flow. If the client needs EXIF preserved for licensing or rights metadata, do the cutout, then in Photoshop or ExifTool copy the EXIF block from the original JPEG onto the PNG before delivery.
Several popular cutout services let you upload an image and get a cutout back via their API. The implicit deal is: you give them training data in exchange for free or cheap processing. That deal is fine for non-sensitive personal photos and disastrous for paid client work under an NDA. Running the model locally in the browser tab is the only architecture that fully sidesteps the question. Your client's pre-launch product photography stays on your laptop. The output PNG is the only thing that ever leaves your machine, and it leaves only when you upload it to your own delivery service.
Drop a JPEG, get a transparent PNG. No signup, no upload, no watermark. Built for photographers who bill hours, not pixels.
Try the Background Remover →