Mirror Fit
Try it yourselfRequest a demo
Virtual fitting for fashion retail

Try it before you
buy it.

Let your customers try on your entire catalogue on their own photo, and mix and match fashion items into complete looks.

The same photo, wearing the bouclé outfit: jacket, blouse and trousersWith Mirror Fit
Original photo: a woman, full body, from the frontYour photo
The KPIs that matter

More conversion, fewer returns, more revenue.

Today's market benchmarks. Once you are live, we measure the same KPIs in your shop and show them in your dashboard.

Conversion1.53%average conversion rate in fashion e-commerceWe measure your uplift against a holdout group, with a confidence interval.
Return rate~90%of all items returned in German e-commerce are fashionWe report the return rate of fitting buyers next to everyone else, clearly labelled as an observation.
Revenue per visitormore revenue per visitorConversion times basket value, minus returns: the number your finance team looks at.
Interactive

Calculate what returns cost you.

Move the sliders to your shop's numbers. Nothing is sent anywhere.
Returns cost you per month€4.5M450,000 returns × €10
you save €200,000 a month, €2.4M a year.Your assumption, not a promise. Once you are live, we measure the real effect.
Sources

IRP Commerce, May 2026 (conversion) · retourenforschung.de, University of Bamberg, 2025 (returns) · EHI Retail Institute, 2025 (calculator start values)

Try it yourself

Try it yourself with your own image.

Choose one of three complete looks and see it on your own photo. Four steps, no sign in.

  1. 1. Look
  2. 2. Photo
  3. 3. Consent
  4. 4. Result
1. Choose a look
Pick a complete look, then upload your own photo. Every piece is tried on at once.
How it works

Four steps. Only what passes is shown.

STEP 1 · ONBOARDINGUnderstand the productA background job reads your catalogue: product photo, worn view, details, sizes and measurements. Not live, so nothing slows down your shop.
STEP 2Upload a photoOne full body photo is enough. Mirror Fit asks for consent first.
STEP 3AICreate the fittingThe garment is placed on the person. Pose and background stay.
STEP 4AICheck against the originalSeveral independent AI judges compare the result with the original and detect invented parts.
CheckedOnly what passes is shownOtherwise a message appears instead of an image that failed the check.
Not passed: new attempt, at most 3
Continuous learningAlways improving
  • New items are read in automatically
  • AI models are updated continuously
  • The workflow keeps getting better
  • We learn from aggregated usage, never from individuals
Measured in our lab

Flawless fits, every time: because your style deserves double the attention.

In this fitting the first attempt cut the coat too short: the hem stopped above the original length. The check rejected it. The second attempt matched the original and was shown.

Dashed line: length of the original coat
Attempt 1: the coat ends above the original lengthAttempt 1 · rejectedAI check
Length: too short. Rejected before anyone saw it.
Attempt 2: the coat reaches the original lengthAttempt 2 · passedAI check
Length: as in the original. Shown to the customer.
82–83%of fittings pass every check on the first attempt
5–9%end without an image instead of a wrong one
25son average from upload to the checked image

Lab tests, September 2026, one person: 150 fittings across 30 items with the improved rules, confirmed by a second run of 45 new fittings. Timing from a smaller run of 3 fittings. Results in your shop will differ; we measure them once you are live.

Retailer dashboard

Your numbers in one place.

Credits with a forecast of when they run out, the quality of every delivered image, and impact measured against a holdout group instead of guessed.

Billed in credits. A standard fitting uses one credit, a complex one more, for example 1.5. Only delivered, checked images use credits. Each month, 25% of unused credits carry over to the next month.

portal.mirror-fit.comSample data, fictional shop
Mirror Fit
Thread / ObjectProfessional plan
OverviewGood morning. Here is your month so far.
Credits in OctoberRuns out around Oct 26
40,500of 50,000 + 1,025 rollover
Forecast at the current pace · buy more, throttle or stop: you decide
Fittings per dayLast 14 days
Quality, 30 days
95.0%delivered
83%first attempt
Add to cart, holdout test+18%
Up next

On our roadmap.

What we are building next, from fit to marketing spend.

Ask about early access
  1. In developmentFit assistantCalculates from the measurements per size how a garment sits on the body.
  2. In preparationCDP connectionPass fitting events and consents to your customer data platform.
  3. PlannedSpend optimisationUse your Google Analytics or other analytics data to see where your marketing budget works best.
  4. PlannedCampaign testingA/B test campaigns on performance and spend against purchase probability.
Technology

A workflow, not a single prompt.

Mirror Fit reads the input in layers, creates the result in layers and checks it in layers.

Multi-layer analysis

Product photo, worn view and item data are analysed separately before anything is created.

Multiple AI judges

Independent checks compare every result with the original.

Loop to perfection

If a check fails, a new attempt starts automatically, at most three times.

Scoring

Every image gets a quality score. Only images above the threshold are shown.

What we check:ButtonsCollarPocketsBeltLengthPatternColourFabricLighting
INTEGRATION

Integration? Don't worry.

A script tag in your online shop or an SDK in your app. Product data is read in the background, and events such as fitting started and result viewed go to GA4 via your Google Tag Manager. No year long engineering project.

Learn more in the technical docs
Data privacy

Your customers' photos stay in good hands.

  • Runs on Google Cloud, on servers in the EU
  • Consent before every upload
  • Photos and results are deleted automatically after 90 days
  • No training on customer photos without separate consent
Request a demo

See a demo based on your collection.

Tell us about your shop. We come back to you with a demo based on your own items.

  1. Tell us about your shopPlatform, catalogue size and the looks you want to see.
  2. We prepare your demoBuilt on your own items, not on stock examples.
  3. We walk you through itLive, with your team, including the dashboard.
Mirror Fit
Virtual fitting for fashion retail

Try it before you buy it.

Let your customers try on your entire catalogue on their own photo, and mix and match fashion items into complete looks.

Request a demoTry it yourself
The same photo, wearing the bouclé outfit: jacket, blouse and trousers
Original photo: a woman, full body, from the frontYour photo
With Mirror Fit
The KPIs that matter

More conversion, fewer returns, more revenue.

Today's market benchmarks. Once you are live, we measure the same KPIs in your shop.

Conversion1.53%average conversion rate in fashion e-commerceWe measure your uplift against a holdout group, with a confidence interval.
Return rate~90%of all items returned in German e-commerce are fashionWe report the return rate of fitting buyers next to everyone else, clearly labelled as an observation.
Revenue per visitormore revenue per visitorConversion times basket value, minus returns: the number your finance team looks at.
InteractiveCalculate what returns cost you.
Returns cost you per month€4.5M
you save €200,000 a month.Your assumption, not a promise. Once you are live, we measure the real effect.
Sources

IRP Commerce, May 2026 (conversion) · retourenforschung.de, University of Bamberg, 2025 (returns) · EHI Retail Institute, 2025 (calculator start values)

Try it yourself

Try it yourself with your own image.

  1. 1 Look
  2. 2 Photo
  3. 3 Consent
  4. 4 Result
Pick a complete look, then upload your own photo. Every piece is tried on at once.
How it works

Four steps. Only what passes is shown.

  1. STEP 1 · ONBOARDINGUnderstand the productA background job reads your catalogue: photos, details, sizes, measurements.
  2. STEP 2Upload a photoOne full body photo is enough. Consent comes first.
  3. STEP 3AICreate the fittingPose and background stay.
  4. STEP 4AICheck against the originalSeveral independent AI judges compare the result with the original.Failed? New attempt at step 3, at most 3
Only what passes is shownOtherwise a message appears instead of an image that failed the check.
CONTINUOUS LEARNINGAlways improving
  • New items are read in automatically
  • AI models are updated continuously
  • The workflow keeps getting better
  • We learn from aggregated usage, never from individuals
Measured in our lab

Flawless fits, every time: because your style deserves double the attention.

Here the first attempt cut the coat too short: the hem stopped above the original length (dashed line). The check rejected it. The second attempt matched the original and was shown.

Attempt 1: the coat ends above the original lengthAttempt 1 · rejected
Length: too short. Rejected.
Attempt 2: the coat reaches the original lengthAttempt 2 · passed
Length: as in the original. Shown.
82–83%pass every check on the first attempt
5–9%end without an image instead of a wrong one
25son average from upload to the checked image

Lab tests, September 2026, one person: 150 fittings across 30 items, confirmed by a second run of 45 new fittings. Timing from a run of 3 fittings. Results in your shop will differ; we measure them once you are live.

Retailer dashboard

Your numbers in one place.

Credits with a forecast, the quality of every delivered image, and impact measured against a holdout group.

Credits in OctoberRuns out ~Oct 26
40,500of 50,000 + 1,025
Quality, 30 days
Delivered95.0%
First attempt83%
Add to cart+18%Holdout
Sample data, fictional shop

Billed in credits. A standard fitting uses one credit, a complex one more, for example 1.5. Only delivered, checked images use credits. Each month, 25% of unused credits carry over to the next month.

Up next

On our roadmap.

  1. In developmentFit assistantCalculates from the measurements per size how a garment sits on the body.
  2. In preparationCDP connectionPass fitting events and consents to your customer data platform.
  3. PlannedSpend optimisationUse your Google Analytics or other analytics data to see where your marketing budget works best.
  4. PlannedCampaign testingA/B test campaigns on performance and spend against purchase probability.
Ask about early access
Technology

A workflow, not a single prompt.

Multi-layer analysisPhoto, worn view and item data, analysed separately.
Multiple AI judgesIndependent checks against the original.
Loop to perfectionFailed check, new attempt, at most three.
ScoringOnly images above the threshold appear.
INTEGRATIONIntegration? Don't worry.Script tag in your shop or SDK in your app. Product data is read in the background, events go to GA4.Learn more in the technical docs
Data privacy

Your customers' photos stay in good hands.

Google Cloud, servers in the EUConsent before every uploadDeleted automatically after 90 daysNo training without separate consent
Request a demo

See a demo based on your collection.

  1. Tell us about your shopPlatform, catalogue size and the looks you want to see.
  2. We prepare your demoBuilt on your own items, not on stock examples.
  3. We walk you through itLive, with your team, including the dashboard.