Our background removal tool processes one image at a time; there's no bulk-upload option for running an entire folder through in a single action. For a small to medium batch, a bit of routine around that one-at-a-time process makes it considerably less tedious than it might first sound.
Setting up before you start
- Gather all the source images into a single folder first, so you're not hunting for files partway through.
- Rename the files with a consistent, sequential naming pattern if they don't already have one, so you can track progress and avoid re-processing the same image twice.
- Create a second, empty folder specifically for the processed results, keeping originals and finished cutouts clearly separated.
A steady one-at-a-time routine
Open the tool, upload the next image, process it, download the result into your results folder, and move to the next one. Working through a consistent, small routine like this is faster in practice than it sounds — the actual processing step is quick, and most of the time is really just the upload-download cycle.
Keeping track of what's done and what's left
For a larger batch, keeping a simple running checklist — even just a plain text list of filenames you tick off as you go — avoids losing track partway through, particularly if you need to pause and come back to the batch later.
Checking each result before moving to the next
A quick glance at each result before moving on — confirming the cutout looks clean and the subject wasn't partially cut off — catches any problem images early, rather than discovering several bad results only after the whole batch is finished. See our guide on why compression can't fix a blurry photo for a related example of why catching a source-quality problem early matters.
When a batch is genuinely too large for this manual approach
For a very large batch — hundreds of images at once — a manual one-at-a-time routine becomes impractical regardless of how efficient the routine is, and a dedicated batch-processing tool or script-based approach becomes the more sensible choice at that scale.
Naming the processed files consistently
Keep the same base filename between the original and the processed cutout (perhaps with a suffix like "-cutout"), so it's always clear which processed result corresponds to which original, particularly useful if you need to reprocess or compare a specific image later.
Doing the batch in one sitting versus spreading it out
For consistency of quality checking, working through a batch in one focused sitting is generally easier to track accurately than spreading it across several separate sessions, where it's easier to lose track of exactly where you left off.
Taking short breaks during a long batch
For a genuinely large batch, a short break every twenty or so images helps you keep spotting quality issues accurately — reviewing results carefully gets noticeably harder once attention starts to fade partway through a long, repetitive session, and a brief pause resets that far more effectively than pushing straight through.
Setting realistic expectations for how long a batch will take
Estimate roughly how long a single image takes through the full upload-process-download-check cycle, then multiply by your batch size to set a realistic time expectation before you start, rather than discovering partway through that the batch is taking considerably longer than assumed and needing to rush the remaining images.
Keeping your results folder clean as you go
Beyond just separating originals from processed results, it helps to periodically glance through the results folder itself and remove or flag any image that needs a redo, rather than letting a growing pile of unreviewed results accumulate until the very end of the batch. Addressing a problem image as soon as you spot it, while the context of what went wrong is still fresh, is generally faster than coming back to it after the whole batch is otherwise finished and you've lost track of exactly what the issue was.
Why a manual routine is often more reliable than it sounds
It's easy to assume a manual, one-at-a-time process must be inherently error-prone or tedious compared to some automated alternative, but for a moderate batch size, a careful manual routine with good habits — consistent naming, a clear checklist, regular quality checks — is often more reliable in practice than it initially sounds, precisely because each image gets individual attention rather than being processed blindly in bulk.
Deciding this is worth doing yourself versus outsourcing it
For a genuinely large, ongoing volume of product photography, it may be worth considering whether this kind of repetitive processing work is better handled by a dedicated team member or outsourced entirely, rather than continuing to absorb it yourself indefinitely. Recognising when a manual routine has outgrown being a reasonable use of your own time is a legitimate business decision, separate from whether the routine itself works well at a smaller scale.
Frequently asked questions
Is there a bulk-upload option for background removal?
Not currently; images are processed one at a time through the tool.
What's the most efficient way to work through a folder of images?
Gather all source files first, use a separate results folder, and work through a steady one-at-a-time routine with a simple checklist for larger batches.
Should I check each result immediately or all at the end?
Check each one as you go, so any problem image is caught early rather than discovered only after the whole batch is done.
What if I have hundreds of images to process?
At that scale, a manual one-at-a-time approach becomes impractical; a dedicated batch tool or script-based approach is more appropriate.
How should I name the processed files?
Keep the same base filename as the original with a clear suffix, so each processed result is easy to match back to its source.
Should I do a large batch in one sitting?
Generally yes, for easier tracking, though pausing with a checklist works fine if needed.
Is the tool free for each image processed?
Yes. No sign-up, no watermark, and files are removed from the server automatically about an hour after processing.