How to remove backgrounds automatically

If you sell on Amazon, eBay, Shopify, or any other marketplace, you already know the drill: every product photo needs a clean, white background. Doing that by hand for dozens or hundreds of images is tedious and expensive.

With a PixelFiddler workflow, you can make it automatic. Upload a raw product photo, and seconds later a marketplace-ready version lands in a separate folder — background removed, white fill applied, saved as a high-quality JPG. No manual editing, no outsourcing.

This guide walks you through building that workflow from scratch.

What you will need

• A PixelFiddler account on any paid plan (background removal uses AI tokens)
• A storage with a folder for raw product photos (e.g., a Hosted storage or an S3 bucket)

AI token usage

Background removal is powered by an AI model and consumes tokens from your plan’s monthly allowance. Each image processed counts toward your token budget. Check your plan’s token limits under Settings before processing large batches.

Step 1: Create a new workflow

1. Open your PixelFiddler dashboard
2. Navigate to Workflows in the left sidebar
3. Click Create Workflow

Give your workflow a clear name so you can identify it later — something like auto-remove-backgrounds works well. You can also add a short description such as “Removes backgrounds from raw product photos and saves clean JPGs.”

Step 2: Configure the trigger

Every workflow starts with a trigger — the event that kicks it off. For this recipe, you want the workflow to run whenever a new photo lands in a specific folder.

1. In the workflow editor, select the File Uploaded trigger
2. Choose the storage where your product photos are stored
3. Set the file pattern to raw-photos/*.jpg

The file pattern tells PixelFiddler to watch only the raw-photos folder and only react to JPG files. This prevents the workflow from firing on unrelated uploads.

Customize the pattern

You can broaden the pattern to catch more file types. Use raw-photos/*.{jpg,png} to include both JPG and PNG uploads, or raw-photos/**/*.jpg to watch subfolders too.

Step 3: Add the Background Removal node

Now add the step that does the heavy lifting.

1. Click the + button after the trigger
2. Search for and select Background Removal
3. Connect the File Uploaded trigger to the Background Removal node by dragging between their connection points

You will see a settings panel for the node. The most important choice here is the AI model:

• BASIC — Faster processing, lower token cost. Good for simple product shots with solid or uniform backgrounds.
• QUALITY — Slower but more precise, especially around fine details like hair, fur, or translucent edges. Uses more tokens per image.

For most standard product photography (items on a table, mannequin shots, flat lays), BASIC handles the job well. Switch to QUALITY when you are shooting jewelry with intricate edges, clothing with loose threads, or anything where the subject blends into the background.

After this step, the image will have a transparent background (the subject is preserved, everything else is removed).

Step 4: Add a Transform Image node

Most marketplaces require a white background, not a transparent one. To convert the transparent PNG output into a white-background JPG, add a transform step.

1. Click the + button after the Background Removal node
2. Select Transform Image
3. Connect Background Removal to Transform Image

Configure the following settings:

• Format: JPG (this automatically replaces transparency with a white fill)
• Quality: 95 (high enough for marketplace standards without bloating file size)

Why 95 quality?

A quality setting of 95 keeps images crisp and detailed while still benefiting from JPG compression. Most marketplaces accept anything above 72, but higher quality makes your listings look more professional and avoids visible compression artifacts around product edges.

Step 5: Add a Save to Storage node

The final step tells PixelFiddler where to put the finished image.

1. Click the + button after the Transform Image node
2. Select Save to Storage
3. Connect Transform Image to Save to Storage

Configure these settings:

• Directory: clean-backgrounds (the folder where processed images will be saved)
• Filename: {orig_name}-clean (appends “-clean” to the original filename so you can tell raw and processed files apart)

For example, if you upload blue-sneaker.jpg to the raw-photos folder, the workflow will produce blue-sneaker-clean.jpg in the clean-backgrounds folder.

Step 6: Save and enable

Your workflow is complete. The full chain looks like this:

File Uploaded —> Background Removal —> Transform Image —> Save to Storage

1. Click Save in the top-right corner of the workflow editor
2. Toggle the workflow to Enabled

The workflow is now live. Every JPG uploaded to your raw-photos folder will be processed automatically.

Step 7: Test it

Upload a product photo to your raw-photos folder to verify everything works:

1. Go to Storage and open your storage
2. Navigate to the raw-photos folder
3. Upload a test image
4. Wait a few seconds, then check the clean-backgrounds folder

You should see a new file with the -clean suffix, featuring your product on a solid white background.

Caution

If the processed image does not appear, check the Workflows log for errors. Common issues include incorrect folder paths, missing storage permissions, or insufficient AI tokens remaining on your plan.

Variation: Keep transparent backgrounds

If you need transparent backgrounds instead of white (for example, for your own website where products float over colored sections), you can simplify the workflow.

Skip the Transform Image node entirely. Connect Background Removal directly to Save to Storage and the output will be saved as a transparent PNG.

File Uploaded —> Background Removal —> Save to Storage

Just make sure the filename pattern preserves the PNG extension, and update your directory if you want these saved separately (e.g., transparent-cutouts).

Variation: Save both versions

Sometimes you need both — a transparent PNG for your website and a white-background JPG for marketplaces. You can handle this in a single workflow using a Broadcast Router.

1. After the Background Removal node, add a Broadcast Router instead of connecting directly to the next step
2. Create two branches from the router:
   • Branch A: Add a Save to Storage node that saves the transparent PNG to a transparent folder
   • Branch B: Add a Transform Image node (format: JPG, quality: 95) followed by a Save to Storage node that saves to a marketplace-ready folder

This way, every upload produces two versions in one pass, using only one set of AI tokens for the background removal.

Next steps

Now that your background removal workflow is running, explore more ways to automate your image pipeline:

• Resize images on upload — Automatically generate multiple sizes for different platforms
• Route images by format — Send PNGs, JPGs, and WebPs to different folders or pipelines
• Background Removal reference — Dive deeper into model options, edge refinement, and advanced settings

Explore all workflow nodes