OnModel
OnModel replaces the models in apparel product photos using AI, letting stores vary ethnicity, age, expression and background, or turn flat-lay shots into on-model images. Delivered as a Shopify app, web app and API.
What is OnModel?
OnModel is an AI product-photography service aimed at online clothing retailers, published by Trendage, Inc., a company that presents itself as supplying software and machine learning to the apparel and footwear industry since 2014.
It covers two distinct products. Model Swap, the mature one, takes a photo that already contains a person and regenerates that person: gender, ethnicity, age, expression and makeup can all be changed while the garment itself stays untouched. It also generates faces on cropped or headless shots, and swaps backgrounds so that a catalogue looks consistent. AI Photoshoot, still in beta and open on application, works the other way round: it takes a flat-lay, ghost-mannequin, hanger or mannequin shot and produces an on-model image from it.
The commercial pitch rests on three arguments the site repeats. Stores using their supplier's stock photos publish the same images as their competitors, which the vendor presents as an SEO handicap. Models that do not match a store's customer demographics cost sales. And a generated image costs a fraction of a studio shoot with models, makeup artists and a photo crew.
Distribution is threefold. A Shopify app, published in February 2023 and rated 4.5 across ten reviews, works from products already loaded in a store. A standalone web app accepts any uploaded image without Shopify. An API, with an iframe SDK that can be styled to the customer's own brand, is offered on request alongside custom exclusive models; the vendor advertises dedicated and serverless GPUs and weekly model updates, but publishes no documentation.
The limits are stated openly. All poses work for model swapping, but AI Photoshoot handles front views only. PNG, JPG and WebP are accepted; RAW and HEIC are not. Jewellery, footwear and child models are listed as roadmap, not features. The vendor also acknowledges that the AI hallucinates and recommends generating three to five images per photo and keeping the best — a workflow expectation rather than a defect, but one worth budgeting for when estimating how far a monthly swap quota will go.
What it does
- Swap the model in an existing product photo, changing gender, ethnicity, age, expression and makeup
- Generate a face on cropped or headless images
- Remove and replace backgrounds to standardise a catalogue
- Turn flat-lay or ghost-mannequin shots into on-model photography
- Process thousands of images in batch and approve them in bulk
- Use your own models instead of the vendor's library
- Embed model swapping into another product through an API or a brandable iframe
When to use OnModel / When not to
A quick filter to help you decide if OnModel is the right fit.
When to use OnModel
- Shopify clothing stores that rely on supplier stock photos and need visuals their competitors do not already publish
- Dropshippers, wholesalers and print-on-demand sellers who cannot fund a traditional photoshoot
- E-commerce and merchandising teams with large catalogues, thanks to batch processing of thousands of images
- Brands wanting their models to match the ethnicity, age or demographics of their customer base
- Product teams and developers embedding model swapping into their own application through the API or iframe
When not to use OnModel
- Jewellery and footwear sellers, categories the vendor lists as roadmap rather than supported
- Brands needing children's or baby models, also announced as future work
- Organisations with documented GDPR obligations, since the site never mentions the regulation, a DPA or its subprocessors
- Companies that must keep their imagery out of AI training, as the terms grant a perpetual training licence with no documented opt-out
- Teams needing back or side on-model shots, or working in RAW or HEIC, none of which are supported
How to use OnModel
A typical end-to-end flow, from setup to results.
- Choose your route: the Shopify app for a store catalogue, the web app for loose images, or the API for an integration
- For Shopify, install the app from the App Store and make sure a product with compatible images is already loaded
- For the web app, create an account from the signup page, then drag and drop an image
- Check the image suits the product: Model Swap needs a photo that already contains a person
- For AI Photoshoot, bulk upload flat-lay, ghost-mannequin, hanger or mannequin shots instead
- Pick a model from the vendor's library or upload your own, then set gender, ethnicity, age, expression and makeup
- Adjust the background and any optional retouching or styling options
- Generate three to five variants per photo, as the vendor recommends, and keep the best result
- For a whole catalogue, switch to batch mode and approve the generated images in bulk
- Publish the retained images back to your store, using the in-dashboard live chat or support form if a result disappoints
Pros & Cons
Pros
- Public, readable pricing with a permanent free tier of 25 swaps and a low paid entry point at 9 USD per month
- Ownership of the generated images is assigned to the user for commercial use, in writing in the terms
- Two independent distribution routes: a one-click Shopify app and a standalone web app that needs no store
- Batch mode genuinely suited to large catalogues, with bulk approval of results
- An API and a brandable iframe for integrators, with the option of exclusive custom models
- A publisher that predates the product and is identifiable: Trendage, Inc., in apparel software since 2014 according to the site
- Unusual candour about the technology's limits, with an explicit roadmap and a recommendation to generate several variants
Cons
- No pricing page on the site itself: rates are only public on the Shopify listing, leaving web app and API pricing undisclosed
- Complete silence on the GDPR, with no data processing agreement, no subprocessor list, no EU representative and no declared hosting location
- A perpetual licence to train machine learning models on uploaded and generated images, with no documented way to opt out
- Aggregated anonymised data may be sold to third parties, and objecting means writing to the publisher
- Two different postal addresses depending on the page, and a privacy policy with no effective date that describes a mobile app the product does not have
- No published support email; support runs through in-dashboard chat and a form only
- A narrow review base of ten Shopify ratings, two of them one-star, reporting failed ethnicity conversions and slow support
Pricing & Plans
A permanent free plan is available, covering 25 model swaps. The cheapest paid plan is Basic at 9.00 USD per month for 300 model swaps per month. All charges are billed in US dollars and collected every 30 days, with no annual billing option offered. These rates are published on the vendor's Shopify App Store listing; the site itself has no pricing page, and neither the standalone web app nor the API carries a publicly stated price.
- Free — 0 USD — 25 model swaps
- Basic — 9 USD per month — 300 model swaps per month
- Growth — 19 USD per month — 1
- 000 model swaps per month
- Advanced — 49 USD per month — 5
- 000 model swaps per month
- API access — price not published
- granted on request
Data, GDPR & hosting
A consolidated view of how OnModel handles your data.
GDPR overview
No implementation is stated. Across every page collected, the GDPR is never mentioned, and neither is the CCPA, a legal basis for processing, data subject rights in the European sense, international transfers, nor an Article 27 representative. Probing the usual paths returned nothing: /gdpr, /dpa, /legal, /security, /trust and /subprocessors are all absent, and no subprocessor list is published. The privacy policy carries no effective date, although it announces that one appears at the top of the page, and it describes itself as covering a mobile application the product does not have. What is offered in practice is access, correction and deletion on written request to a Privacy Compliance contact, plus a stated minimum age of thirteen. The compliance flag is negative because the site is silent, not because it refuses compliance.
Who owns the data?
The terms of use draw a two-way split. Users own the assets they generate: Trendage, Inc. assigns them all rights, title and interest for commercial use. In exchange, using the service grants the publisher a perpetual, worldwide, non-exclusive licence over the actions, prompts and images fed into the tool and over the assets it produces. That licence covers use, reproduction, modification, publication, sublicensing, distribution and the training of machine learning models. It is granted without payment, the terms stating that use of the service is itself sufficient consideration. Users may not upload images they hold no rights to, nor generate clothing models under eighteen.
Reuse rights
Generated images can be reused freely and commercially without asking permission: the terms assign all rights, title and interest in them to the user, and the FAQ repeats the point. Access to the service itself is licensed more narrowly, being non-exclusive, limited, non-transferable and revocable, but ownership of the output does not depend on holding a paid plan. The reverse direction is broader than most. Whatever a user uploads or generates may be reused by Trendage, Inc. to provide, maintain, promote and improve the service, to comply with the law, and to train machine learning models. No mechanism to exclude images from that training is documented anywhere on the site.
Data retention & training
Hosting summary
The site declares nothing about data hosting. No country and no region is named in the privacy policy, in the terms or anywhere else, and there is no trust or security page: the usual paths for one do not exist. The API page advertises dedicated and serverless GPUs that scale with demand, but names neither a cloud provider nor a location. The only element touching storage is the privacy policy's description of an encrypted user database, with no further detail. Outside the vendor's own statements, a technical check resolves the domain to an Amazon Web Services address geolocated in the United States. That is an observation about where the website is served, not a declaration about where customer images or account data are processed, and it should not be read as one. Buyers with a hosting requirement will have to obtain the answer directly from the publisher, because the site does not provide it.
Things to keep in mind
Risks and trade-offs to weigh before adopting OnModel.
- The terms grant a perpetual, worldwide licence to train machine learning models on every image you upload or generate, without payment and with no documented opt-out
- The privacy policy allows aggregated anonymised data to be sold to third parties, and objecting requires writing to the publisher
- Nothing is published about the GDPR, a data processing agreement, subprocessors or where data is hosted, which makes vendor due diligence hard for regulated buyers
- Generated models are synthetic people: using them without disclosure can mislead shoppers about fit and appearance, and may create returns rather than sales
- Two conflicting postal addresses appear on the site, and the Delaware one is a widely shared registered-agent address rather than an operating office
- The domain expires on 02/10/2026, shortly after this review, which is worth re-checking before committing a whole catalogue to the tool
- Editorial upkeep is irregular: several pages carry wrong titles and meta descriptions, and a plan name quoted in the FAQ no longer matches the live pricing
Setup & Integrations
Technical difficulty
Low for the two main routes. The Shopify app installs in one click and only requires a product already loaded with compatible images; the web app needs an account and a drag-and-drop. Neither calls for technical skill. Expect iteration rather than difficulty: the vendor recommends generating three to five variants per photo and keeping the best. The API is a different matter, since access is granted on request through a form and no public documentation exists, so an integration requires a prior exchange with the publisher and development resources.
Deployment
Integrations
Supported languages
Behind OnModel
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
What does OnModel actually do?
Do I need a Shopify store to use it?
Who owns the images it generates?
Are my images used to train the AI?
What does it cost?
Is there an API?
Which product categories are not supported?
Is the tool GDPR compliant?
Who publishes OnModel?
Should you pick OnModel?
OnModel is a shipping product rather than a promise: the Shopify app has been published since February 2023, is billed on four public tiers, and carries ten dated reviews spanning 2023 to 2025. The publisher, Trendage, Inc., predates the tool and is named consistently in the terms of use and in Shopify's developer record. For a clothing store working from supplier stock photos, the proposition is concrete: unique imagery, models matched to a customer base, and a free tier of 25 swaps that lets anyone judge output quality before paying 9 USD a month.
The reservations concern transparency, not existence. No pricing page lives on the site itself, so web app and API rates stay unknown. The site is entirely silent on the GDPR, publishes no data processing agreement and names no subprocessors, while the terms grant a perpetual licence to train machine learning models on uploaded and generated images with no opt-out documented. Two different postal addresses appear depending on the page, and the privacy policy carries no effective date.
Maturity is uneven: Model Swap is in production, AI Photoshoot is beta and front-pose only, and jewellery, footwear and child models remain roadmap items. Ten ratings are too few to settle the question of output quality, though two of them report failed ethnicity conversions and slow support.
The sensible path is to spend the free tier on the garments that matter most, and to treat the training licence as the real decision point: for a store publishing its own catalogue it is usually acceptable, for a brand under contractual image restrictions it may not be.
- Choosing a selection results in a full page refresh.
- Opens in a new window.