AI Photo Cleanup Studio

Problem

Use this when an existing photo contains an object that must be removed, such as a bystander in a travel photo, a vehicle in a street scene, an extra hand or wire in a product image, or clutter in a room photo. The goal is to remove only the selected object while preserving the person, product, text, and background.

How It Works

• Establish the base image and reuse the uploaded artifact.
• Define what should disappear, what must remain, and what should fill the gap.
• Use edit_regions with rectangles or full-image editing; rectangles are safer near faces, products, or text, while scattered items may be handled across the whole image.
• Use beatra.models.list to inspect live model capabilities instead of assuming controls or limits.
• Confirm the route, count, canvas size, total runs, and maximum cost before payment, and keep a stable client_request_id for each call.
• Submit each approved beatra.images.edit call once, record the task ID, and poll with beatra.tasks.get until it reaches a terminal state.
• Deliver the actual dimensions, MIME type, file size, artifact, resolved model, usage, and billing.

Boundaries

It is suitable for object removal and surrounding inpainting, not for full style transfer, realism enhancement, professional headshots, product background replacement, generation without a source image, or video retouching. The covered area is newly generated from surrounding context, not recovered original detail; if it originally contained a face, price, serial number, or document that needs accuracy, warn the user before submission. Multiple targets are split into separate paid runs, and after a failed run, later rectangles must be reconfirmed on the new base image.