Framekit templates
Start from a designer-made template
Use template
Use template
Use template
Use template
Use template
Use template
Use template
Use template
Removing the background from an image is a solved problem. Every tool in this comparison runs a segmentation model, every one of them produces a PNG with an alpha channel, and in our view the differences on a clean product shot are now small.
The prices are not. The same thousand cutouts cost $0.00 from one tool on this page and $200.00 from another, and the expensive one is the one most people have heard of.
A background remover is a tool that detects the subject in an image and separates it from everything behind it, returning a transparent PNG or a replaced background.
It is used most heavily by anyone selling physical products online, where a catalogue of a thousand items means a thousand cutouts, and where a per-image price quietly becomes a line in the budget.
Cut-outs are rarely the end of the job.
For most people on this page the finished PNG goes onto a portfolio, a case study or a product listing, and Framekit is one place that page can live: hosting on a global CDN and a digital-download store, on your own domain from the $9 Starter plan rather than a marketplace profile.
The best background remover in 2026 depends entirely on volume.
For occasional use, Clipdrop's free tier gives you 20 removals every 24 hours at no cost. For anything programmatic, BRIA at $0.018 an image and Photoroom at $0.02 are the cheapest credible APIs.
For genuinely high volume, rembg is free, open source under an MIT licence and runs on your own machine with no per-image cost at all. remove.bg is the best-known name here and at roughly $0.20 an image it is about eleven times the price of the cheapest API for output that is not eleven times better.
Updated in September 2026: we re-verified Framekit's plans and store fees and corrected our own section, which previously implied a Framekit store could carry physical goods; it sells digital downloads only.
Background remover prices are as read in August 2026.
On 12 October 2026 we withdrew a 40-image test set and the quality findings drawn from it, which we could not document, so output claims now rest on how these models work and on a test you run yourself.
How We Compared 10 Background Removers
We compared ten background removers on what each vendor publishes: per-image, per-credit and per-megapixel prices, free allowances, batch and API options, and where processing happens, read on each vendor's own pricing page on 9 August 2026, then converted everything to the cost of 1,000 cutouts.
We did not run an image set through the tools for this article, so nothing below is a measured edge-quality result.
Findings are labelled by how we got them. Prices were read off each vendor's own pricing page on 9 August 2026, with the exceptions noted in individual entries.
What we say about edges, hair and semi-transparent material is our editorial reading of how this class of model works rather than a result from a test set, and the fifteen-minute test near the end is how to judge it on your own images.
Anything about support or long-term reliability is reported and marked as such.
Two currency notes matter here more than in most comparisons. Erase.bg and Slazzer both served us euro pricing from a European address, so their figures below are euros and labelled.
Pixian prices by megapixel rather than per image, so a single per-image number does not exist for it and we give the range instead.
In one lineten tools converted to the cost of 1,000 cutouts, every price read at the vendor's own page on 9 August 2026 or explicitly labelled, and output quality left to a test you run on your own hardest images.
What the Prices Show Against a Converged Technology
The technology in this category has largely converged and the pricing has not. That single sentence is the whole comparison, and it is why the ranking below is ordered by cost far more than by output.
- The same 1,000 cutouts cost between $0.00 and about $270, or about $200 at remove.bg's standard credit rate, depending only on which tool you opened.
- 7 of the 10 publish an API, so most of the field can be wired into a catalogue pipeline rather than used one image at a time.
- Hair and semi-transparent fabric are where differences between segmentation models still show, so they are the images to try before you pay anyone.
- The cheapest credible API is roughly 11 times cheaper per image than the best-known consumer service.
- 3 of the 10 can process unlimited images at no marginal cost, either by running locally or through a one-time purchase.
- Glasses and fine jewellery chain are the hard cases for this whole class of model, so budget manual cleanup for them whichever tool you pick.
For the pick matched to your volume rather than the whole field, jump to the decision guide.
What 1,000 Cutouts Actually Cost
Background removal is bought by the thousand or it is bought once, and almost every pricing page is designed to obscure which of those you are doing. Converting every model to the cost of 1,000 finished cutouts is the only way to see the category clearly.
| Tool | How it is priced | Cost per image | Cost per 1,000 cutouts |
|---|---|---|---|
| rembg | Free, open source, runs locally | $0.00 | $0.00 |
| Clipdrop free tier | 20 removals every 24 hours | $0.00 | $0.00 up to about 600 a month |
| PhotoScissors Desktop | One-time purchase, unlimited | $0.00 after purchase | About $19.99 to $39.99 once |
| Pixian.ai | By megapixel | $0.0009 to $0.0185 | About $0.90 to $18.50 |
| Erase.bg | Credit packs | About 0.01 euro at 1,000 credits | About 14.54 euro |
| BRIA | Pay as you go | $0.018 | $18.00 |
| Photoroom API | Pay as you go | $0.02 | $20.00 |
| Pixelcut Pro | Subscription | Not metered per image | $8 to $10 a month |
| Slazzer | Credit subscription | About 0.10 euro at 1,000 | About 98.70 euro a month |
| remove.bg | Credits | About $0.16 to $0.27 | About $160 to $270 |
The distance between the top and the bottom of that table is the point. A shop with a thousand product images pays nothing to run rembg on a laptop overnight, $18 to call BRIA's API, or up to $270 to put the same images through the best-known service in the category.
None of that would matter if the expensive one were dramatically better, and the section below on shared segmentation models explains why that is unlikely; your own hardest images are how to check.
In one linethe same thousand cutouts range from $0.00 to about $270, while most of these tools run comparable segmentation models, so price no longer tells you much about quality.
The 10 Best Background Removers in 2026: Full Comparison
Scores are our rating out of 10, based on published features and prices, across four weighted criteria: documented help with hard subjects such as hair, glass and mesh, including refinement brushes and shadow generation (30%), cost per image at realistic volume (30%), how it fits into an actual workflow including batch processing and API access (25%), and ease of getting started (15%).
Cost carries an unusually high 30% here because the core segmentation is now largely shared: when most tools run comparable models, price stops being a tiebreak and becomes the decision.
| Tool | Type | Price | Batch | Best for | Score |
|---|---|---|---|---|---|
| rembg | Open source, local | Free | Yes | Volume with zero marginal cost | 9.0 |
| Photoroom | App and API | $0.02 per image API | Yes | Product sellers who want both | 8.7 |
| BRIA | API | $0.018 per image | Yes | Cheapest credible API | 8.4 |
| Clipdrop | Web app and API | Free tier, 20 per 24h | Limited | Occasional use at no cost | 8.1 |
| Pixian.ai | Web and API | By megapixel, from $0.0009 | Yes | Cheapest at small image sizes | 7.8 |
| PhotoScissors | Desktop | About $19.99 to $39.99 once | Yes | Offline, unlimited, bought once | 7.5 |
| Pixelcut | App | $10 a month, $8 annual | Yes | Sellers who want templates too | 7.1 |
| Erase.bg | Web and API | About 0.01 euro at volume | Yes | Cheap credit packs | 6.8 |
| Slazzer | Web, desktop, plugin | About 0.10 euro per image | Yes | Photoshop integration | 6.4 |
| remove.bg | Web and API | About $0.16 to $0.27 | Yes | Nothing we would pay this for | 6.0 |
remove.bg finishing last is the clearest verdict in this comparison and it is not a quality judgment, because it is a well-regarded product with a well-documented API.
It is ranked last because it costs roughly eleven times what a comparable API costs, and in a category built on shared models that is the main thing left to judge.
1. rembg: Free, Local, and Unlimited
rembg is an MIT-licensed open-source library that runs the same class of segmentation model as the paid services, on your own machine, with no account, no API key and no per-image cost.
How it worksmore like a utility than a product, and once it is set up it disappears into a folder-watching script. Point it at a directory of 500 product shots, come back, and there are 500 transparent PNGs.
Because it wraps the same published model families many paid services build on, there is little reason to expect clean product shots to come out worse, and hair is the place to compare it with a paid API on your own files.
The real number$0.00, at any volume, forever. Against remove.bg at roughly $0.20 an image, a catalogue of 5,000 products is $1,000 saved. Against BRIA it is $90 saved on the same catalogue, which is the more honest comparison and still free.
The detail nobody writes aboutbecause it runs locally, your images never leave your machine.
For anyone under a client NDA, working on unreleased products, or photographing people who did not consent to their picture being uploaded to a third party, that is not a nice-to-have, it is the only option on this page that satisfies the requirement.
The gotchasetup is the price. It is a Python library, and while desktop wrappers exist, the first installation involves a runtime, a model download and the usual dependency friction. Budget an hour if you are comfortable with a terminal and an afternoon if you are not.
Skip it ifyou need one cutout today and have never opened a command line, in which case Clipdrop's free tier does it in a browser in ten seconds.
Verdict: the correct answer for anyone processing images at volume and the wrong one for anyone processing three. Free, private, unlimited, and it costs you a setup afternoon exactly once.
2. Photoroom: Best Combination of App and API
Photoroom is the strongest all-rounder here because it is genuinely two products: a polished app for people who want to drag a folder in, and an API for people who want to wire it into a catalogue pipeline.
How it worksit is built for product photography specifically, which shows in the details. It generates contact shadows on cutout products, a step most tools here leave to you, and the batch flow assumes you have a folder rather than an image.
The real numberthe Remove Background API is $0.02 an image and the Image Editing API is $0.10, with 10 free production calls on account creation and 1,000 monthly sandbox calls that come back watermarked.
The app subscription runs from about $7.50 a month billed annually.
The detail nobody writes aboutcalling the wrong API costs you five times as much. If you are on the Basic plan and make an Image Editing call, you are charged the Plus price of $0.10 rather than $0.02, and the vendor states this plainly on its own pricing page. On a 10,000-image run that mistake is $800.
The gotchacredit and export allowances on the app plans vary by country, which Photoroom states in its own FAQ. The plan described in a US comparison is not necessarily the plan you will be sold.
Skip it ifyou only need the API and price is the only axis, where BRIA is 10% cheaper for the same operation.
Verdict: the best choice for a product seller who wants one vendor for both the manual work and the automated work. The shadow generation alone justifies it over a raw API for anyone photographing objects.
Selling those products yourself? We compared the platforms and their cuts in the best website builders to sell digital products.
3. BRIA: The Cheapest Credible API
BRIA is an enterprise-oriented visual AI platform whose background removal endpoint happens to be the cheapest per-image price from a serious vendor in this comparison.
How it worksas an API rather than an app. There is a web interface, and it is not why you are here.
The real numberbackground removal is $0.018 an image on pay-as-you-go with no generation cap, and the free tier includes 100 generations with no card required.
That is $18.00 per thousand cutouts, against $20.00 for Photoroom and roughly $200 for remove.bg at standard credit rates.
The detail nobody writes aboutBRIA publishes per-operation pricing across its whole catalogue rather than bundling everything into credits, so background removal at $0.018, eraser at $0.02 and generative fill at $0.03 are all separately visible.
In a category built on opaque credit conversions, being able to read the actual unit price is genuinely unusual.
The gotchait is aimed at companies, and the surrounding product is a platform rather than a tool. If you want a button in a browser, you are buying a lot of machinery you will not use.
Skip it ifyou are not writing code. This is an integration, not an application.
Verdict: the price leader among paid APIs and the right default for anyone building background removal into a product or a catalogue pipeline. At $0.018 an image the argument against it is not cost.
4. Clipdrop: The Best Free Tier in the Category
Clipdrop, from the Stability ecosystem, offers the most generous free allowance here by a distance, and for a large number of people it means this category costs nothing at all.
How it worksa clean browser tool with a set of adjacent features, cleanup, relighting, uncrop, that are genuinely useful when a cutout is not quite the whole job.
The real numberthe free plan includes 20 background removals every 24 hours, alongside unlimited text-to-image and uncrop, 20 daily upscales and 50 daily text removals.
Twenty a day is roughly 600 a month at no cost, which covers a lot of small shops entirely.
The detail nobody writes aboutthe free allowance resets on a rolling 24-hour basis rather than monthly, which changes how you use it. A shop with 600 products cannot do them in a weekend and can do 20 a day for a month, and for a business with no budget that is a real workflow rather than a teaser.
The gotchathe Pro plan's price did not render on the pricing page when we read it, showing a placeholder rather than a figure, so we are not quoting one. The free tier is the reason to be here anyway.
Skip it ifyou need more than 20 a day, at which point the free tier becomes a bottleneck and rembg becomes the obvious answer.
Verdict: the best zero-cost option that requires no installation, and the tool to send someone who needs a few cutouts and no complexity.

5. Pixian.ai: Cheapest If Your Images Are Small
Pixian prices by megapixel rather than per image, which is the most honest pricing model in this comparison and also the hardest to compare against everyone else.
How it worksstripped back. There is very little product around the function, which suits it.
The real numberan image costs between 0.05 and 1 credit depending on its megapixel count, working out to roughly $0.0009 for the smallest files and about $0.0185 for a 25-megapixel one.
Credit packs run from about $5 for 250 up to roughly $1,455 for 100,000. The free tier is limited to 0.25 megapixels with low-priority processing, which is a preview rather than a usable output.
Pixian's page served us euro pricing, so the dollar figures here are converted and approximate.
The detail nobody writes aboutmegapixel pricing rewards you for not sending unnecessary resolution.
A web store needs a 1,500-pixel product image, not a 6,000-pixel one, and downscaling before removal rather than after can cut your bill by more than an order of magnitude on this platform. No per-image vendor gives you that lever.
The gotchathe flipside is that a catalogue of large raw files gets expensive quietly, because the price you sanity-checked on a small test image is not the price you pay on production files.
Skip it ifyour images are large and you cannot downscale them first, where a flat per-image API is more predictable.
Verdict: the cheapest option in this comparison for anyone processing web-sized images, and a pricing model more vendors should copy. Resize first, then remove.
6. PhotoScissors: Buy It Once, Run It Forever
PhotoScissors is a desktop application sold for a one-time fee, which in a category dominated by per-image billing makes it structurally different from almost everything else here.
How it worksconventional desktop software with a manual refinement brush, which matters more than it sounds: painting a correction in the same window, without exporting to another editor, is what makes the hard cases manageable compared with the pure web tools.
The real numberreported at about $19.99 for Desktop Standard and about $39.99 for Desktop Pro as one-time purchases, with unlimited processing under the licence. Sources disagree, with one listing the desktop version at about $29.99, so treat the exact figure as reported rather than confirmed.
The gotchaa perpetual licence in this category means a model frozen at purchase time. Segmentation quality has improved noticeably year on year, and software bought once does not get the next model unless the vendor ships an update.
Skip it ifyou want the current state of the art, which is what per-image services sell you alongside the removal.
Verdict: the right answer for someone processing images offline in bulk on a fixed budget, and worth checking which version the price you are quoted actually buys.
7. Pixelcut: Background Removal Inside a Seller's Toolkit
Pixelcut aims at people selling on marketplaces, wrapping removal in templates, product shots and generative backgrounds rather than selling the operation on its own.
How it worksphone-first and template-driven. For a marketplace seller who needs a listing image rather than a transparent PNG, the workflow is shorter than in any pure removal tool because it ends at a finished image.
The real numberPro is $10 a month or $8 a month billed annually, and Business is $30 a month or $24 annually.
The free plan includes limited background removal, limited upscaling, watermark-free export and 300 daily credits for the generative features.
The gotcha"limited background removal" on the free tier is not quantified on the pricing page, which makes it hard to plan around. Watermark-free export on a free plan is genuinely generous and the ambiguity about volume is not.
Skip it ifyou want a transparent PNG rather than a finished listing image, where cheaper tools do the same job.
Verdict: good value at $8 a month for a marketplace seller, and the wrong shape of product for anyone who just needs cutouts.
8. Erase.bg: Cheap Credits, Awkward Currency
Erase.bg, from the Pixelbin team, sells credit packs at prices that fall sharply with volume and undercut most of the field at the top tier.
How it worksstraightforward web removal with batch processing on paid plans.
The real numbercredits are sold at about 3.64 euro for 10, about 7.27 euro for 100, and about 14.54 euro for 1,000, which works out to roughly 0.01 euro an image at the top pack. Three free credits arrive on sign-up.
The page served euro pricing while the FAQ references a $5 entry point, so the currency you are shown may depend on where you are.
The detail nobody writes aboutcredits expire after one year. Buying the 1,000-credit pack for the unit price and then using 300 of them means you paid roughly three times the headline rate, which is the standard trap in credit-based pricing and one this vendor states clearly rather than burying.
The gotchathe price per image swings by a factor of 36 between the smallest and largest pack, so buying the wrong pack size costs more than choosing the wrong vendor.
Skip it ifyour volume is unpredictable, where pay-as-you-go APIs with no expiry are safer.
Verdict: genuinely cheap at volume and only if you actually use the volume. Work out your annual image count before choosing a pack.
9. Slazzer: The Photoshop Integration
Slazzer's differentiator is not its model, it is that it ships desktop applications for Windows, Mac and Linux plus a Photoshop extension, so removal happens where you already work.
How it worksthe Photoshop extension is the reason to consider it. Removing a background without leaving the document you are compositing in saves a round trip that adds up across a shoot, and no other tool here does it as directly.
The real numbermonthly credit subscriptions run from about 11.55 euro for 100 credits up to about 1,811 euro for 50,000, with the per-image rate falling from about 0.12 euro to about 0.04 euro.
The 1,000-credit tier is about 98.70 euro a month, roughly 0.10 euro an image. Two free credits are included, with unlimited free previews on the website. Slazzer served us euro pricing from a European address.
The gotchaat roughly 0.10 euro an image on the mid tier, Slazzer is about five times the price of BRIA, and nothing in the shared-model argument below suggests five times better output. You are paying for the integration.
Skip it ifyou do not live in Photoshop, which is the only thing here that justifies the premium.
Verdict: worth it for a retoucher who works in Photoshop all day and difficult to justify for anyone else.
10. remove.bg: The Famous One, and the Most Expensive
remove.bg largely created this category and is still the first result most people find. Its output is good. Its pricing has not moved with the market.
How it worksa polished service with the best-documented API in the comparison and integrations everywhere. Its long-standing reputation is on hard cases such as curly hair, which is worth checking against a cheaper API on your own images before you pay its rate.
The real numbercredits are reported at about $0.20 each at standard rates, with subscription plans reported from about $9 a month for 40 credits, roughly $0.27 an image at the entry tier, falling to about $0.16 an image on larger plans.
A pay-as-you-go starter pack of 3 credits for $3 works out to $1.00 an image. remove.bg's pricing page did not render its figures to us, so all of these are reported rather than read at source.
The detail nobody writes aboutthe entry subscription has a worse per-image rate than the mid tiers, at roughly $0.27 against about $0.16, so the cheapest plan is the most expensive way to buy. That shape is common in credit pricing and it catches the smallest customers hardest.
The gotchaat roughly $0.20 an image you are paying about eleven times BRIA's rate. Competitors now advertise themselves explicitly as being multiples cheaper, which is an unusual position for a category leader and a fair summary of where it sits.
Skip it ifcost matters at all, which at any volume above a few dozen images it does.
Verdict: a good product at a price the market moved away from. If you are already integrated and the volume is small, staying is defensible. If you are choosing today, the same job costs a tenth as much.
Why All These Tools Suddenly Produce the Same Result
In one linethe outputs converged because most of this category is built on a small number of shared open-source segmentation architectures, which is why a free local tool matches a paid API and why price stopped tracking quality.
This is the mechanism behind the central finding of the comparison, and it is worth understanding before you pay anyone.
Background removal is an image segmentation problem: for every pixel, decide whether it belongs to the subject or to the background, and for edge pixels decide what proportion of each.
The research on this has been public and rapid, and a handful of model families, developed openly and published with permissive licences, became good enough that the remaining differences are small.
That has three consequences that show up directly in the pricing table above.
A free local tool can match a paid service because it may be running a descendant of the same published architecture. rembg is explicit about this: it is a wrapper around pretrained models rather than a proprietary system, which is exactly why it costs nothing and performs respectably.
The differentiators moved somewhere else. Vendors that are genuinely ahead are ahead on the surrounding work rather than the segmentation itself: shadow generation, batch tooling, refinement brushes, integrations, colour handling on the matte edge.
Photoroom's product shadows are a better reason to choose it than its cutouts are.
Price stopped signalling quality. In a category where the core operation is commoditised, an eleven-times price difference is a statement about market position and integration cost rather than about output.
That is not an accusation of anything; it is what happens to any product whose core function becomes freely available.
The practical consequence for a buyer is simple and slightly uncomfortable for the vendors.
Test the cheapest credible option on your own hardest images first, because the burden of proof has flipped: the question is no longer whether a cheap tool is good enough, it is whether an expensive one is meaningfully better on your specific images, and only your own images can answer that.
Batch Processing Is What Actually Decides Your Afternoon
In one linethe difference between a tool that processes a folder and one that processes an image is the difference between a background task and a lost afternoon, and it is the feature least visible on pricing pages.
At one image, everything here is equivalent. At 400, the workflow is the entire product.
| Tool | Batch method | Practical ceiling |
|---|---|---|
| rembg | Command line over a folder | Whatever your disk holds |
| BRIA | API, no generation cap | Whatever you can queue |
| Photoroom | App batch plus API | Large, API-limited |
| Pixian | Web batch and API | Credit balance |
| Erase.bg | Batch on paid plans | Credit balance |
| Slazzer | Desktop app and API | Credit balance |
| PhotoScissors | Desktop batch | Unlimited under licence |
| Pixelcut | App batch | Plan limits, unstated |
| Clipdrop | One at a time on free tier | 20 per 24 hours |
| remove.bg | API and integrations | Credit balance |
Two things fall out of this that the price table does not capture.
Clipdrop's free tier, generous as it is, is a single-image workflow. Twenty a day is a real allowance and it is twenty separate interactions, which for a catalogue is not a workflow at all.
And the tools with no per-image cost, rembg and PhotoScissors, are the only two where you can point the software at everything you own and walk away without watching a meter.
For anyone with a genuine archive, that psychological difference turns out to matter as much as the money.
What to Do With the PNG Afterwards
In one linea transparent PNG is an intermediate rather than a deliverable, and three decisions after the cutout affect your page more than the choice of remover did.
Do not ship PNG to the web. A transparent PNG of a product photograph is frequently several times the size of an equivalent WebP with an alpha channel, and page weight is a real cost on a product catalogue. Convert after removal, keep the PNG as the master.
Put the shadow back. A cutout product floating on white looks cut out, and the single thing that makes an e-commerce image read as photographed rather than processed is a plausible contact shadow.
Photoroom generates these, most tools do not, and doing it manually in any editor takes under a minute once you know to.
Check the matte edge against your actual background. A cutout that looks clean against the checkerboard preview can show a fringe of the original background colour when placed on a coloured page.
Look at your PNG on the exact background it will sit on before approving a batch of 400, because fixing that afterwards means running the whole batch again.
One more, specific to catalogues: keep your originals. Every one of these tools will improve, and the model that handles your glassware badly today will handle it better in two years. A folder of cutouts with no sources is a decision you cannot revisit.
Three Tools That Did Not Make the Ten
Exclusions are usually more informative than another near-identical entry, so here is what was set aside and why.
Adobe Express and Canva both remove backgrounds competently and are covered in the section on software you already own rather than as ranked entries.
Both vendors also block automated access to their pricing pages, so neither could be priced at source alongside the rest.
Cutout.pro was excluded because its pricing page returned a 404 when we checked it, and a comparison built on verified prices cannot include a vendor whose prices we could not reach.
The large field of single-purpose web removers was excluded as a group.
Many are thin interfaces over the same open models discussed above, they compete on advertising rather than capability, and the honest recommendation for anyone considering one is rembg, which is the same technology with no upload and no cost.
Where These Tools Still Fail
In one linefour subject types remain hard for every background remover in this comparison including the most expensive, and knowing them tells you which images to budget manual time for rather than which vendor to switch to.
The marketing images for every one of these tools show a person against a plain wall or a shoe on white. Here is what happens outside that.
Hair against a busy background. Fine strands need per-pixel transparency rather than a hard edge, and every tool here approximates it.
The good ones produce soft, plausible edges; the weaker ones produce a halo of background colour or a hard cut through the hair.
In our view it is the largest quality differentiator left in the category, and nothing in the published pricing suggests it tracks price.
Glass, mesh and anything semi-transparent. A wine glass or a mesh fabric should carry partial alpha, showing some of whatever is behind it.
Most tools treat the subject as opaque and hand you a glass with a solid interior, which for product photography means manual work regardless of which vendor you paid.
Fine repeating structure. Jewellery chain, bicycle spokes, wicker and lace all break the same way, with the gaps between the structure filled in. Expect any of these tools to lose some of a fine chain.
Low contrast between subject and background. A grey jumper on a grey wall gives the model very little to work with, and the failure is unpredictable rather than uniform: sometimes an arm disappears, sometimes the wall comes along.
The practical consequence is a workflow rather than a purchase.
Sort your images into easy and hard before you start, run the easy ones through whichever tool is cheapest, and reserve manual masking or a refinement brush for the small proportion that needs it.
Paying a premium vendor for the whole batch to handle the 5% better is the expensive way to solve a cheap problem.
The Cost That Is Not on Any Pricing Page
In one lineat any realistic volume your own time is a larger line item than every tool on this page combined, which is why the correct question is how many images need a human afterwards rather than what each removal costs.
Run the arithmetic that the pricing pages leave out.
A thousand product images at BRIA's $0.018 costs $18.00. If 10% of those need a minute of manual cleanup, that is 100 minutes, and at even a modest $40 an hour of your time that cleanup costs $66.67, nearly four times the software.
At remove.bg's roughly $0.20 the software becomes $200 and the labour is unchanged, so the tool choice moves the total from $85 to $267 while the cleanup rate moves it far more.
That reframes the whole comparison. A tool that is $0.002 cheaper per image and needs twice as much manual correction is the expensive option.
This is the argument for testing on your own hard images rather than trusting any ranking including this one, because the cleanup rate is a property of your photographs rather than of the software.
Three things reduce that rate more than switching vendors does.
Shoot for the cutout. A product photographed against a background that contrasts with it, lit so the edges are defined, segments almost perfectly in every tool here.
Fifteen minutes of care at the shoot removes hours of cleanup afterwards, and it is the highest-return habit in this entire workflow.
Sort before you batch. Separating the easy 90% from the hard 10% before processing means you can run the bulk through the cheapest option and give the difficult images the attention they need, rather than paying a premium across the whole set to slightly improve a minority.
Fix the source rather than the output. If a whole product line segments badly, the cause is usually the lighting or the surface it was shot on, and re-shooting twenty items is often faster than manually masking two hundred images of them.
There is also a fourth option that belongs in an honest comparison: outsourcing. Manual clipping-path services still exist and can produce better edges on genuinely hard subjects than an automated tool, because a person is drawing the path.
For glassware, jewellery and anything with complex transparency, that remains a real answer, and it is priced per image by the service rather than by any vendor here, so compare it against your own cleanup time rather than against a subscription.
The Remover You Already Own
In one lineif you have Photoshop, Affinity, Pixelmator or Canva already, you have a background remover with no per-image cost, and for low volume that is often the correct answer instead of any tool on this page.
This section exists because a lot of people arrive at this question without checking what is already installed.
Affinity removes backgrounds through its selection and refinement tools, and since Canva made the suite free on Mac and Windows in October 2025 that capability costs nothing at all. It is manual rather than one-click, and it is the most controllable option here.
Photoshop has had a one-click Remove Background action for years, and its Select Subject has improved substantially. If you are already paying for Creative Cloud, per-image services are a duplicate purchase for anything but bulk.
Pixelmator Pro, a $49.99 one-time purchase on a Mac, includes a machine-learning background removal inside the editor, at no extra cost per image.
Canva removes backgrounds on its paid tiers, which is how a very large number of people already do this without thinking of it as a separate category.
The honest rule: below about 50 images a month, use what you own. Above that, the time you spend clicking becomes the real cost and a batch tool pays for itself immediately, at which point the question is which of the ten above rather than whether.

Run This Test Before You Pay For Anything
Take the five hardest images you own. Not your best product shot, your worst: something with hair, something semi-transparent, something with fine repeating detail, and something where the subject and background are similar colours.
Run all five through Clipdrop's free tier, which costs nothing and needs no account setup, and look at the results at 100% rather than as thumbnails. Then ask one question: how many of the five need manual cleanup regardless.
If the answer is none, you have just discovered that the free tier solves your problem and this article is finished for you. If the answer is one or two, that is the normal result and your decision is purely about volume and price.
If the answer is four or five, no tool on this page will fix that and your real answer is a refinement brush and ten minutes per image, which makes PhotoScissors or an editor you already own the right purchase rather than any API.
That test takes fifteen minutes and it is more informative than every comparison table above, because it uses the images you actually have.
Where Cut-Out Product Images Earn Their Keep
Every tool above produces a transparent PNG. None of them is where that PNG makes money, which for most people reading this is a product page. That is what we build, so read this as disclosure rather than an eleventh entry.
Framekit is an AI website builder for photographers, designers and creators, with a built-in store for digital downloads, so the images you have just cut out can go onto portfolio, case-study and product pages you own rather than into a marketplace listing that ranks you against nine competitors selling the same thing.
For a product photographer the practical benefit is that the portfolio and the site around it are one system; a shop selling physical goods still needs a commerce platform, because Framekit sells digital files only.
Two honest limits. Framekit does not remove backgrounds and does not compete with anything on this page.
And selling through Framekit carries a transaction fee that starts at 5% on the free and Starter plans, drops to 3% on Pro at $19 a month and reaches 0% on Business at $39 a month, with Stripe processing on top on every plan and the flat-versus-percentage crossover against the free plan at around $780 a month in sales.
The Decision Guide
Volume decides this almost entirely, so work down these questions in order and stop at the first yes rather than comparing edge quality between tools whose edges are nearly identical.
Do you already own Photoshop, Affinity, Pixelmator or a Canva subscription, and do you need fewer than about 50 cutouts a month? If yes, use what you have. Every tool on this page is a duplicate purchase at that volume.
Do you need fewer than 20 a day and want nothing installed? If yes, Clipdrop's free tier, at no cost, in a browser.
Are you processing hundreds or thousands of images and comfortable with a terminal? If yes, rembg. Free, unlimited, local, and your images never leave your machine.
Do your images fall under an NDA, or show people who have not consented to third-party processing? If yes, rembg is the only option here that keeps them on your own hardware.
Are you wiring this into a product or a catalogue pipeline? If yes, BRIA at $0.018 an image, or Photoroom at $0.02 if you also want its product shadows and app.
Are your images web-sized rather than full-resolution? If yes, Pixian, where megapixel pricing makes small files dramatically cheaper. Downscale before you remove, not after.
Do you live in Photoshop and value not leaving it? If yes, Slazzer, and accept that you are paying roughly five times BRIA's rate for the integration.
Do you sell on marketplaces and want a finished listing image rather than a PNG? If yes, Pixelcut at $8 a month billed annually.
Frequently Asked Questions
What is the best background remover in 2026?
The best background remover in 2026 depends on volume rather than quality, because most of these tools now run comparable segmentation models, so on clean product shots price rather than output separates them.
For occasional use Clipdrop's free tier gives 20 removals every 24 hours at no cost, for programmatic use BRIA at $0.018 an image is the cheapest credible API, and for high volume rembg is free, open source and runs on your own machine with no per-image cost.
Is there a genuinely free background remover?
Yes, and there are two kinds. rembg is MIT-licensed open-source software that runs locally with no account, no limit and no cost at any volume.
Clipdrop's free tier allows 20 background removals every 24 hours in a browser with nothing installed, which is roughly 600 a month.
Beyond those, most "free" tools in this category deliver a low-resolution preview and charge for the usable file.
How much does background removal cost per image?
Background removal costs between about $0.0009 and about $0.27 per image depending entirely on the vendor.
BRIA is $0.018, Photoroom's API is $0.02, Pixian ranges from about $0.0009 to $0.0185 by megapixel, Slazzer is about 0.10 euro, and remove.bg is reported at roughly $0.16 to $0.27.
For 1,000 cutouts that is a range from $0.00 to about $270 for output that is largely equivalent.
Is remove.bg worth the price?
remove.bg is a well-regarded service with a long-standing reputation on hard cases such as curly hair, and at roughly $0.20 an image it costs about eleven times what BRIA charges for the same operation.
If you are already integrated and processing small volumes, the switching cost may exceed the saving.
If you are choosing today at any meaningful volume, the price difference is difficult to justify, and competitors now advertise themselves explicitly as multiples cheaper.
What is the cheapest way to remove backgrounds from 1,000 product photos?
The cheapest way to remove backgrounds from 1,000 product photos is rembg, which is free, open source and runs locally with no per-image cost, so the only expense is an hour of setup.
Among paid options, BRIA at $0.018 an image totals $18.00 for a thousand and Photoroom's API at $0.02 totals $20.00.
Downscaling images to web size before processing also cuts the bill substantially on megapixel-priced services such as Pixian.
Do background removers work on hair?
Background removers handle hair better than they did two years ago and it remains the hardest common case, because fine strands need partial transparency rather than a hard edge.
Hair is where differences between models still show, and nothing in the published pricing suggests they track price. Curly hair against a busy background is where the weakest tools produce a coloured halo or cut straight through strands.
Can I remove backgrounds without uploading my images anywhere?
Yes. rembg runs entirely on your own machine, and desktop applications such as PhotoScissors process locally after installation.
This matters for images under an NDA, unreleased products, and photographs of people who have not consented to third-party processing.
Every web-based tool in this comparison, including the free tiers, involves uploading your image to a company's servers.
Which background remover has an API?
Photoroom, BRIA, remove.bg, Slazzer, Erase.bg, Pixian and Clipdrop all offer APIs, and the price differences between them are large.
BRIA is $0.018 an image and Photoroom is $0.02 for background removal, while remove.bg is reported at roughly $0.20.
Photoroom charges $0.10 rather than $0.02 if you call its Image Editing endpoint rather than its Remove Background endpoint, which is an easy and expensive mistake at volume.
Do I need a paid tool if I already have Photoshop or Canva?
Below roughly 50 images a month, no. Photoshop's one-click Remove Background, Affinity's selection tools, Pixelmator Pro's machine-learning removal and Canva's paid-tier removal all handle the common cases without a separate subscription.
Above that volume the time spent clicking becomes the real cost, and a batch tool or an API pays for itself quickly.
Why does the same tool charge different prices for different images?
Some vendors price by megapixel rather than per image, so a 25-megapixel file costs many times what a web-sized one does. Pixian is the clearest example, charging between 0.05 and 1 credit depending on image size.
This is arguably the fairest model, because processing cost genuinely scales with pixel count, and it means resizing before removal rather than after can reduce a bill by more than an order of magnitude.
Do background removal credits expire?
Often, and it is the most common way people overpay in this category.
Erase.bg states that its credits are valid for one year, and buying a large pack for the low unit price then using a third of it means paying roughly three times the headline rate.
Pay-as-you-go APIs with no expiry are safer for unpredictable volume, and subscription credits that reset monthly are the least forgiving of all.
What is the best background remover for e-commerce product photos?
Photoroom is the strongest choice for e-commerce specifically, because it generates realistic contact shadows on cutout products, which is the step most sellers do manually afterwards, and it offers both a batch app and a $0.02 API.
Pixelcut is a reasonable alternative for marketplace sellers who want a finished listing image rather than a transparent PNG, at $8 a month billed annually.
Can these tools handle transparent objects like glass?
Poorly, across the whole category. A wine glass or a bottle needs partial alpha so that some of the background shows through, and most tools treat the subject as opaque and return a solid interior.
Expect that across price tiers, including the most expensive service. For product photography involving glass, budget manual masking time regardless of which vendor you choose.
Is rembg hard to install?
rembg is a Python library, so the first installation involves a runtime, a model download and the usual dependency handling, which is roughly an hour for someone comfortable with a terminal and an afternoon for someone who is not.
Desktop wrappers exist that package it with a graphical interface and remove most of that friction. Once installed it processes folders unattended with no per-image cost at any volume.
Final Verdict: The Best Background Remover in 2026
There is no single answer here, and pretending otherwise would misread a category whose core technology has converged. The quality gap has largely closed and the price gap has not, so the honest recommendation is by volume.
Under 50 images a month, use the remover in software you already own. Under 20 a day, Clipdrop's free tier costs nothing. Above that, rembg is free, unlimited, private and worth an afternoon of setup.
If you need an API, BRIA at $0.018 an image is the price leader and Photoroom at $0.02 is worth the 10% premium for product shadows and a proper app.
Where this whole category loses, plainly: hair against busy backgrounds, glass, mesh and fine repeating detail remain hard for every tool here including the most expensive one, and no vendor claims to have solved them.
If your images are mostly hard cases, the answer is a refinement brush and manual time rather than a different subscription.
Where Framekit loses, and it is worth stating directly: we do not remove backgrounds, we are not an alternative to anything ranked above, and if that is what you came for then the ten tools here are your list and we are not on it.
Our store also sells digital downloads only, so a catalogue of physical goods needs a commerce platform that handles stock, shipping and fulfilment, none of which we do.
The short version. Free and unlimited: rembg. Free and effortless: Clipdrop, 20 a day. Cheapest API: BRIA at $0.018. Best for product sellers: Photoroom. Cheapest for web-sized images: Pixian. Buy it once and work offline: PhotoScissors.
Already in Photoshop all day: Slazzer. Best-known and hardest to justify: remove.bg.
Related readingthe best Photoshop alternatives, the best AI tools for photographers, the best website builder for product photographers, how to start selling digital products, and the best AI image upscalers.


