Ghost Mannequin · Capability
Ghost Mannequin API — Photta
Strip the mannequin out of a garment photo and keep interior construction — seams, linings, tags — fully intact. Hollow-man or flat-lay finish, batch-ready, 4 credits at 2K.
In one sentence
Photta's ghost mannequin API takes a garment-on-mannequin photo and returns an invisible-mannequin image with the interior of the garment reconstructed. POST to `/api/v1/ghost-mannequin` with a `product_image_url`, a `style` (hollow_man or flat_lay), and a resolution. Poll `/api/v1/ghost-mannequin/:id` every 3 seconds until status flips to `completed`.
curl -X POST https://ai.photta.app/api/v1/ghost-mannequin \
-H "Authorization: Bearer $PHOTTA_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"product_image_url": "https://example.com/tshirt-on-model.jpg",
"style": "hollow_man",
"resolution": "2K"
}'What to expect
Typical completion
1–3min
2K / 4K credits
4 / 6
Styles
2
Batch-ready
yes
Code, end to end
Pick your language
Same REST contract; the ghost mannequin endpoint also supports batch processing for bulk catalog runs.
How it works
One photo in, one invisible-mannequin photo out
Upload the source, pick a style, submit, poll, download.
- 01
Step 1
Generate an API key
Ghost mannequin uses the same credentials as every other Photta capability — one key, one credit balance.
- 02
Step 2
Upload the source photo
You can pass a public URL in `product_image_url` or upload the bytes first via `/api/v1/uploads`. The endpoint accepts JPG, PNG and WebP.
- 03
Step 3
POST the ghost mannequin job
POST to `/api/v1/ghost-mannequin` with `product_image_url`, `style` (hollow_man or flat_lay), `resolution`. Pass `webhook_url` if you want a ping when the job completes — polling still works as the primary path.
- 04
Step 4
Poll until the job completes
GET `/api/v1/ghost-mannequin/:id` every 3 seconds. Ghost mannequin is faster than try-on — usually 1 to 3 minutes — because no new person is being rendered.
- 05
Step 5
Download and persist
The output is a PNG with the mannequin removed and the interior of the garment reconstructed. Ideal for PDP ghost shots, 3D-feel product galleries and flat-lay inserts.
Questions other developers ask
Questions developers ask before shipping ghost mannequin
What is ghost mannequin photography?+
A PDP treatment where the garment appears to float on an invisible body, showing shape and interior construction at once. Historically it's stitched together from multiple studio shots; Photta's API produces it in a single call from one on-mannequin photo.
How much does ghost mannequin cost?+
4 credits at 2K, 6 credits at 4K — one credit cheaper at each tier than try-on. Failed jobs refund automatically.
What's the difference between hollow-man and flat-lay style?+
Hollow-man keeps the 3D body shape of the mannequin — the garment retains volume, shoulders and sleeves read naturally, and the interior is reconstructed. Flat-lay collapses the garment to a clean 2D lay-out, ideal for capsule collections or marketplace listings that expect a consistent front-facing shape.
Can I batch-process multiple photos?+
Yes. POST to `/api/v1/ghost-mannequin/batch` with an array of source images. The response gives you a batch ID plus per-item generation IDs. Poll `/api/v1/ghost-mannequin/batch/:id` for an aggregate status or each item individually.
Does it preserve interior seams and labels?+
Yes — the model reconstructs the inside of the garment (neckline interior, cuff interior, waistband lining) rather than hallucinating a generic surface. Tags and printed labels visible on the original photo are retained when possible.
How do I poll for the result?+
GET `/api/v1/ghost-mannequin/:id` every 3 seconds. Batch jobs poll a parent endpoint that returns an aggregate status plus a per-item breakdown.
Ghost Mannequin · Capability
Create an account and get an API key
Photta's ghost mannequin API takes a garment-on-mannequin photo and returns an invisible-mannequin image with the interior of the garment reconstructed. POST to `/api/v1/ghost-mannequin` with a `product_image_url`, a `style` (hollow_man or flat_lay), and a resolution. Poll `/api/v1/ghost-mannequin/:id` every 3 seconds until status flips to `completed`.