Marketplace main-image rules, transparent PNG workflows, and downstream AI generators all benefit when the product subject is isolated from clutter. Rhinobird provides a dedicated matting flow on AI Background Removal. This article covers upload requirements, the Generate step with credit display, quality checks on hairline edges, and how cutouts feed AI E-commerce Image Sets and Agent Chat skills.
Open AI Background Removal
From AI Tools, open AI Background Removal (matting workspace). The interface centers on a required Upload Image control and a Generate action that shows credit cost. Locate History for downloads and Login / Sign up if you are testing while signed out.
Choose source photos that matting handles well
Matting works on studio shots, phone photos, and moderately busy scenes. Improve outcomes with:
- Contrast between product edge and background when possible.
- Even exposure on the SKU; crushed shadows lose fine detail after cutout.
- Minimal motion blur on labels you must keep readable post-export.
Extreme fine strands (fur, lace) may need manual touch-up afterward via Image Editor; matting still saves time versus manual pen tool work for standard packaging.
Upload and Generate
Add the file under Upload Image. Review credits on Generate, then press the button. Complete Login / Sign up if a dialog appears, and run Generate again on the same page without re-uploading if the session preserved your file.
Processing time depends on resolution. When finished, open History, preview the cutout on a neutral checker or white backdrop mentally, and download the transparent or cleaned subject your pipeline expects.
Validate cutout quality before reuse
- Edge halos — light fringing around dark bottles; regenerate or edit if halos show on white main images.
- Missing handles or spouts — retake with clearer separation or edit prompts in Image Editor.
- Label integrity — matting should not crop mandatory regulatory text; if it does, use a wider source photo.
Feed downstream PDP and ad workflows
Use the cutout as input to AI E-commerce Image Sets when lifestyle modules previously picked up background clutter from the original JPEG. Pass the same asset into Background Generation or Model Handheld so compositing focuses on the SKU, not the original wall or table.
For PDP long-form modules via AI E-commerce PDP, a cleaner upload reduces ghost edges in close-up modules.
Failure states
- Generate disabled — confirm an image is uploaded.
- Subject eaten into — use higher-resolution source with more padding around the product.
- Semi-transparent areas wrong — glass bottles may need a second pass or Image Editor fix on highlights.
- Credit retry confusion — check History before paying for duplicate runs on the same upload.
AI Background Removal is the fast hygiene step that preserves subject fidelity for listings and every later AI scene—upload once, validate edges, then route the cutout into sets, PDP tools, or Agent Chat skills without carrying original background noise forward.
Store the cutout alongside your original JPEG in your asset library with matching filenames so merchandisers always know which master fed a given listing main image. That pairing speeds reruns when packaging art changes mid-season.
If your channel requires pure white mains, place the downloaded subject on white in your CMS or design tool after matting rather than asking matting alone to invent studio lighting you cannot control in post.