Virtual try-on / MERCHANT WORKFLOW
Should You Preserve a Fashion Model's Pose or Generate a New Model Scene?
Use FASHN Try-On Max with a model image when the merchant wants to retain an existing person's pose and background while changing the garment. Use Product to Model when the merchant wants a newly generated person-and-product image guided by product photography, optional references, and a prompt. FASHN explicitly distinguishes these endpoints. A face reference in Product to Model does not make it an exact-pose clothing-replacement operation.
Choose the output contract before selecting references
FASHN's Product to Model API documentation states that merchants who want to combine a garment with an existing model photograph while preserving its pose and background should use Try-On Max with model_image.
Product to Model has a different contract. It takes a product image and generates a fashion image. It can accept an optional face reference, inspiration image, and background reference, but it does not replace clothing while retaining the pose and background of the face-reference photograph.
The difference is operational. A merchant may have a licensed studio model photograph whose body position and framing have already been approved. That is a preservation task. Another merchant may have only a garment photograph and want a new campaign pose. That is a scene-generation task.
| Requirement | FASHN route | Primary review |
|---|---|---|
| Retain existing model composition | Try-On Max | Pose, background, garment substitution |
| Create a new model scene | Product to Model | Generated pose, product fidelity, identity guidance |
Neither route proves measured garment fit or guarantees flawless anatomy.
Prepare the appropriate authority for each route
For pose preservation, supply an authorized model image and the correct garment source. The model photograph governs the person and composition, while the garment image governs the merchandise.
For Product to Model, the required input is product_image. The optional image_prompt is an inspiration image supplied as an image URL or data URI; it is not a text prompt. It guides pose, environment, and lighting. Written appearance or styling instructions belong in the separate prompt string. Optional face_reference is also an image input and guides identity. The documented face_reference_mode can favor the reference face's pose and expression or the generated composition, but this is not equivalent to preserving the full original model photograph.
Product to Model also supports background_reference. If both an inspiration image and background reference are supplied, FASHN documents different responsibilities for the person's appearance and the environment.
Confirm the actual endpoint, account eligibility, input restrictions, and current quote before submission. Product to Model is documented as a preview-lifecycle endpoint, so do not assume every account or production environment treats it as an unrestricted stable service.
Hypothetical jacket campaign with two different briefs
Hypothetical exercise. No FASHN API request or generation was performed. The fictional jacket is JACKET-OLIVE-08, documented by jacket-olive-front-v4.png and a matching construction-detail photograph.
Brief POSE-HOLD-01 has authorized model photograph model-standing-studio-r2.jpg. The model's standing position and gray background must remain the intended composition. The proposed endpoint is Try-On Max.
Brief POSE-NEW-02 needs a walking pose in an outdoor scene. The merchant supplies the same jacket source and proposes Product to Model with prompt adult fashion model walking outdoors, olive jacket visible, unobstructed front closure. The optional face reference is face-campaign-r1.jpg, used only with appropriate authorization.
| Brief | Composition authority | Endpoint | Review condition |
|---|---|---|---|
| POSE-HOLD-01 | Existing standing photograph | Try-On Max | Pose and background remain appropriate |
| POSE-NEW-02 | New scene brief | Product to Model | Walking action and garment remain credible |
These are proposed inputs, not observed model outputs.
Diagnose a changed pose differently from a wrong jacket
If the first route changes the standing pose substantially, inspect whether the correct endpoint and model image were used. Do not assume a Product to Model face reference is a substitute for Try-On Max's model-image contract.
If the second route generates an acceptable walking pose but changes the jacket's pocket construction, reject the garment depiction and return to the product evidence. A new pose does not authorize a new jacket design.
If the face resembles the reference but the background differs, check whether the brief actually requested scene generation and whether the intended background reference was supplied. Identity guidance and background ownership are separate controls.
If a pose hides the jacket's required front closure, choose a clearer pose or add an authentic detail view. The occlusion guide
CartVisual's fashion-image workflow is separate from FASHN's endpoints and does not expose their API controls by implication.
Use the example with your own product records
The filled records are hypothetical editorial examples. The entry-template row is for your own files, settings and observations; it does not connect to another application.
| Brief | Garment SKU | Garment source | Model or face reference | Endpoint | Pose requirement | Background requirement | Actual result | Next action |
|---|---|---|---|---|---|---|---|---|
| POSE-HOLD-01 | JACKET-OLIVE-08 | jacket-olive-front-v4.png | model-standing-studio-r2.jpg | Try-On Max | Preserve standing pose | Preserve studio setting | Not generated | Verify source composition and garment |
| POSE-NEW-02 | JACKET-OLIVE-08 | jacket-olive-front-v4.png | face-campaign-r1.jpg | Product to Model | Generate walking pose | Outdoor scene | Not generated | Review new pose, identity guidance, and product fidelity |
What the sources establish
- FASHN — Product to Model API: Explicitly distinguishes Product to Model from Try-On Max pose and background preservation, and documents product, face, inspiration, background, and generation inputs.
- FASHN — Getting Started with Virtual Try-On: Provides vendor examples of garment replacement, retained model composition, and separate later pose changes.
Sources and further reading
Recorded source-review dates are listed individually below. An undated source does not inherit the article’s publication date. Software and platform rules can change; check the current original before acting.
- FASHN — Product to Model API · Review date not separately recorded
- FASHN — Getting Started with Virtual Try-On · Review date not separately recorded
Frequently asked questions
Will Product to Model preserve the pose of my face-reference photograph?
FASHN explicitly says Product to Model generates a new image rather than retaining the reference photo's pose and background. Use the documented Try-On Max route when preserving an existing model image is the requirement.
Can a generated walking image establish how the jacket fits during movement?
No. It is a visual creative candidate, not a measured garment-fit or motion test.
YOUR NEXT CAMPAIGN
Create assets from a clear source
Bring your reviewed source photo or brand brief to CartVisual. Check the current tools, account access and credits before generating, then compare every result with your source.
Create a CartVisual account Open the workspace Check access and credits