---
title: "Fix product photos without leaving the batch"
description: "Remove a background, upscale, expand, edit with a prompt or put a garment on a model, right from a batch row. The fixed photo fills that row."
url: https://kalpana.one/features/ai-image-editing
---

# Fix product photos without leaving the batch

AI image editing. Remove a background, upscale, expand, edit with a prompt or put a garment on a model, right from a batch row. The fixed photo fills that row.
Availability: live.

## Fix a photo without leaving the batch
Click a row's image, then Edit on any photo. Every tool opens right there, and the result fills that row as soon as it is ready, saved to your library as a new image. No other app, and nothing to download and upload again.
- From the batch grid, a campaign feed or the template editor
- The original photo stays as it was
- Crop, resize and colour expansion are free

## Remove the background
Keep the product or the model and make everything behind it transparent. The cutout is saved as a new PNG, the same size as the photo, ready for a clean template.
- A transparent PNG, same size as the source
- JPEG, PNG or WebP photos
- Running it again on the same photo is free

## Upscale small photos
Supplier photos are often too small for a story or for marketplace zoom. Upscale makes a sharper copy at twice or four times the size.
- 2x or 4x, with the output size shown first
- Up to 8192 by 8192 pixels in area
- JPEG and PNG photos

## Expand to a new ratio
A square product shot becomes a story or a banner. Generative expand fills the new space to match the scene.
- Square, story, landscape, portrait or a custom size
- Up to 5000 by 5000 pixels
- An optional prompt for the new background

## Edit with a prompt
Change the setting, the light or the background by describing it. Add up to three reference images to guide the result.
- Keeps the photo's shape, or one of nine ratios
- Up to three reference images
- Saved back to your library as a new image

## Put your garment on a model
Virtual try-on dresses the person in a photo in up to three garments, such as a shirt and trousers. A styling instruction asks it to keep the pose, face and background.
- One to three garment photos
- Flat lays or product shots on plain backgrounds work best
- Keeps the model photo's shape, or one of nine ratios

## Extend with a flat colour
For a product shot on a plain background, Color expand grows the canvas right in the browser and fills the new space with a solid colour or a gradient. No AI involved.
- Solid colour or gradient, with an eyedropper to sample the photo
- Square, story, landscape, portrait or a custom size
- Save a copy to the library or download it

## Capabilities
- Remove background: A transparent PNG cutout of the product or model.
- Upscale: A sharper copy at 2x or 4x for zoom, print and large templates.
- Generative expand: Extend images to fit 9:16, 16:9 and more.
- Prompt edits: Describe the change you want, with up to three reference images.
- Virtual try-on: Dress a model photo in up to three of your garments.
- Color expand: Free in the browser: grow the canvas to a new ratio and fill the new space with a solid colour or a gradient, no AI needed.
- Crop and resize: Free in the browser, to a preset ratio, exact pixels and PNG, JPEG or WebP.
- In templates and batches: Run any tool from the image picker and the result fills that layer or row.

## FAQ

**Do I have to leave the batch to fix a photo?**
No. In the batch grid or a campaign feed, click a row's image, then Edit on a photo, and pick a tool. When the edit is done, the new image fills that row; the original stays in your library.

**How is AI editing billed?**
AI tools use AI credits, which are separate from render credits and are bought in packs. Removing a background, upscaling and expanding cost one AI credit each; a prompt edit or a virtual try-on costs two. Crop, resize and colour expand are free.

**Is my original photo changed?**
No. Every tool saves its result to your asset library as a new image and leaves the original as it is.

**Can I run a tool on many photos at once?**
Not yet. Each tool works on one photo at a time, from the library or from the image picker in a template or a batch. Running the same tool on the same photo with the same settings again reuses the earlier result for free.

**Does virtual try-on keep the model's face and pose?**
It is asked to, through the styling instruction, and it usually keeps the pose and framing. It redraws the whole photo, though, so compare the face and hands with the original before you use the result in an ad.