Best Virtual Try-On APIs for Developers

The best virtual try-on API for developers is the one whose request flow, garment coverage and extras match what you are building. Most try-on APIs take a person image and a product image and return a generated image, so the real differences are authentication, sync or async handling, and whether sizing or a storefront widget comes with it.

ZeDaq.ai publishes this blog and is one of the APIs below. Facts about the other providers come only from their own documentation or websites, linked in place and checked November 2026.

If you haven't built with a try-on API before, our try-on API guide for e-commerce developers explains the basic request and response pattern. This article compares the providers.

What should a developer compare between try-on APIs?

Five things decide how much work an integration takes:

1. Authentication. A header key is simplest. Check whether keys can be restricted to your domain, because browser code exposes them.
2. Request flow. Some APIs answer in one call. Others return a job ID that you poll, or call your webhook when the image is ready.
3. Inputs. Base64 images, image URLs, or both. Check size and format rules.
4. Garment coverage. Clothing only, or also shoes, bags, jewelry and glasses.
5. Extras. A ready-made widget, size recommendation or image editing can save you weeks.

Glasses are a special case. Most eyewear try-on runs on the live camera, not on a single photo, so it often needs a different SDK. Our comparison of eyewear virtual try-on software covers those vendors.

Which virtual try-on APIs are worth testing?

ZeDaq.ai Try-On API

The ZeDaq.ai API reference documents POST /api/v1/tryon, authenticated with an x-api-key header. You send the shopper photo and the product image as base64, plus an optional garment_type of upper_body, lower_body or dresses.

The response is either the finished image or a 202 with a job_id and a poll_url that you check until the render is ready. The docs say usage is only charged once a render succeeds. Keys can be locked to your domains, and requests from other origins are refused.

Two extras set it apart. The same account gives you "Find my size" body-scan sizing matched to the store's own size chart, and a script-tag widget if you'd rather not build a front end.

FASHN API

The FASHN API fundamentals page shows authentication with an Authorization: Bearer header. You POST to https://api.fashn.ai/v1/run with a model_name and inputs, get a prediction ID back, then poll /v1/status or set up a webhook.

Its docs home says the API covers image generation, virtual try-on, model creation and editing. It names Try-On Max as the recommended try-on endpoint for most uses, with Try-On v1.6 kept for real-time, cost-sensitive work. The setup guide says API credits are bought on demand.

Google Vertex AI Virtual Try-On

Google Cloud's Virtual Try-On docs describe the virtual-try-on-001 model. You provide an image of a person and an image of a clothing product, and it generates an image of the person wearing that product. A sampleCount setting returns between one and four images per request.

It is part of Vertex AI on Google Cloud, so it may suit teams already on that platform. The page describes a model, not a storefront product, so plan to build the shopper interface and any sizing yourself.

Perfect Corp YouCam AI API

Perfect Corp's AI API page lists a range of try-on APIs. Under fashion it names clothes, fabric, bag, scarf, shoes and hat try-on, and it also lists jewelry, watch, makeup and hair try-on. It offers a free API key with free credits and describes a pay-per-use model for fashion.

The page didn't list an eyewear frame try-on API when checked, and it showed no unit prices, so check its pricing page before you plan costs.

How do the APIs compare side by side?

	ZeDaq.ai	FASHN	Vertex AI Virtual Try-On	Perfect Corp AI API
Auth	x-api-key header	Authorization: Bearer	Not stated on the try-on page	Free API key offered
Flow	Sync or 202 plus polling	Job ID plus polling or webhook	One to four images per request	Not stated on overview page
Garments	Upper body, lower body, dresses	Try-on plus image tools	Clothing products	Clothes, shoes, bags, hats and more
Sizing	Body-scan sizing on your chart	Not mentioned	Not mentioned	Not mentioned
Storefront widget	Script tag and Shopify app	Not mentioned	No	Not mentioned
Domain-locked keys	Yes	Not mentioned	Not mentioned	Not mentioned

Sources, checked November 2026: ZeDaq.ai API reference, FASHN docs, Google Cloud docs and Perfect Corp AI API. "Not mentioned" means the page we checked didn't say, not that the feature is missing.

What does a first request look like?

Here is a minimal ZeDaq.ai call, so you can see the shape most try-on APIs share:

curl -X POST https://www.zedaq.ai/api/v1/tryon \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "person_image_base64": "<base64 photo>",
    "product_image_base64": "<base64 product image>",
    "garment_type": "dresses"
  }'

Handle both outcomes in your code:

1. A finished image. Show it to the shopper.
2. A pending job. Read the poll_url, check it every couple of seconds, and stop when the image comes back.
3. An error. Show a friendly retry message, and log the error code for monthly cap or rate-limit issues.

For a full walkthrough with a mobile front end, see adding try-on to your own fashion app.

How do you pick the right API for your project?

Match the API to the job, not to a ranking.

• You run a fashion store and want try-on plus sizing. Pick an API that includes both, or you will stitch together two vendors and two shopper flows.
• You want try-on now and may skip the front-end build. Favor a provider with a ready widget you can drop in while you build the custom version.
• You also need marketing images and AI models. A provider with image generation and editing in the same API may save a second contract.
• You are already on Google Cloud. A model inside that platform keeps billing, auth and logging in one place.
• You sell accessories like bags, shoes or hats. Check which garment types each API names before you commit.

A worked example: two projects, two picks

Project A is a developer adding try-on to a client's dresses store on WooCommerce. The client also wants a size answer on each product page and doesn't want a long build. An API that comes with body-scan sizing and a script-tag widget, such as ZeDaq.ai, covers both needs while a custom front end is planned.

Project B is an in-house team at a retailer already running on Google Cloud. It wants to generate try-on images inside its own pipeline and has designers to build the interface. A model inside Vertex AI keeps auth, billing and logs in the platform it already uses.

Neither project needs the "best" API overall. Each needs the one that removes the most work for its own job.

A testing checklist

Run every candidate through the same tests:

• Five product photos: a print, a dark garment, a loose fit, a dress and trousers.
• Three person photos with different lighting and poses.
• Time from request to finished image, measured on your own network.
• Error handling when you send a blurry or cropped photo.
• Clear billing: what counts as one call, and whether failed renders are charged.
• Key security: can you restrict the key, and what happens if it leaks?

What about body scanning instead of try-on?

Try-on and body scanning answer different questions. Try-on shows how the garment looks on the shopper. Body scanning measures the shopper so you can recommend a size.

Some providers do one, some do both. If you are weighing a scan-first or 3D approach, our comparison of try-on and 3D body scanning explains the trade-offs.

FAQs

Do try-on APIs work with any e-commerce platform?

Yes, in most cases. A REST API works from any back end that can make HTTPS requests. You still need to build the shopper interface, unless the provider also offers a widget.

Should I call a try-on API from the browser?

Only if the key can be locked to your domain. Otherwise, call the API from your server so the key stays private.

Why do some try-on APIs return a job ID instead of an image?

Generating an image can take several seconds. A job ID lets your app stay responsive while it polls or waits for a webhook.

Do I need a separate API for size recommendation?

It depends on the provider. Of the pages we checked, only ZeDaq.ai mentions body-scan sizing alongside try-on, so with the others you may need a second sizing tool.

Can I test a try-on API for free?

Several providers offer a free tier or free credits. ZeDaq.ai has a Free plan with 3 try-ons a month, and paid plans start at $14 a month with a 3-day trial.