Getting started

New Diffusion is a fashion AI platform. Open Playground (Workflows) to run tools. The left menu matches the groups below: Create, Edit, Studios, Image Tools, and Video.

1Sign in and open Playground from the top navigation.
2Pick a tool from the left menu — each card includes a How to use panel.
3Upload references, set options, Generate — results appear in preview and Generation.

Group	Tools	Jump to
Create	Create image, Create image with Twin model, Create Avatar
Edit	Edit Image
Studios	Fashion E-commerce Studio, Fashion Creative Studio, Twin Model E-commerce Studio, Twin Model Creative Studio, Packshot Studio, Lookbook Generator
Image Tools	Resize/Upscale, Fix Logo, Remove Background, Refine Face, Change hairstyle, 360° view generator, Style transfer, Twin Model Depth Pose Transfer, Twin Model Pose Transfer, Prompt Builder
Video	Video

All workflows

Full reference for every live tool — same names as the Playground menu. Use the table of contents to jump quickly.

Create

CreateZ-Image Turbo

Create image

Basic text-to-image. Type a prompt and generate — no Twin model or LoRA required.

Best for: Moodboards, concepts, backgrounds, and quick fashion ideas from text only.
Inputs: Text prompt; aspect ratio; megapixel; batch 1–6.
Outputs: Still images in Generation and Gallery.

How to use

1Write a clear text prompt describing the image you want.
2Choose aspect ratio (and megapixel if needed).
3Pick how many images to generate (1–6).
4Generate — Z-Image Turbo creates the image (no Twin / LoRA).

Tips

Be specific: subject, fabric, lighting, camera, mood.
Start at 2:3 or 1:1 for catalogue-style frames.

Open in Playground

CreateZ-Image Turbo + Twin LoRA

Create image with Twin model

Text-to-image locked to a Twin Model face from the library.

Best for: Consistent identity shots of a trained Twin without garments or scenery uploads.
Inputs: Prompt + Twin Model (required); strength; aspect; batch.
Outputs: Images with Twin identity applied.

How to use

1Write a prompt describing the look you want.
2Open Select Twin Model and pick one Twin face.
3Adjust model strength if needed, then set aspect ratio.
4Generate — result appears in center preview and Generation.

Tips

Include the Twin in Select Twin Model before generating.
Keep strength around 0.7–0.9 unless you need a softer likeness.

Open in Playground

CreateNano Banana Pro

Create Avatar

Generate clean portrait references for the Avatar Library.

Best for: Building new avatar faces (name, ref, gender, age) for Fashion studios.
Inputs: Description; gender; age range; aspect; batch.
Outputs: Portrait images; Super Admin can save to Avatar Library.

How to use

1Describe the avatar face you want (features, hair, age feel).
2Set gender and age range.
3Choose aspect ratio and how many images to generate.
4Generate — Super Admin can save the best result to Avatar Library.

Tips

Ask for front-facing head-and-shoulders, soft studio light, plain background.
Avoid props or hands covering the face.

Open in Playground

Edit

EditNano Banana Pro

Edit Image

Multi-image composition / text+image editing (up to 10 references). No Twin model.

Best for: Combining references, guided edits, and creative composites.
Inputs: 1–10 images + prompt; optional makeup / lighting / shadows / pose mode.
Outputs: Edited or composed stills.

How to use

1Upload one or more reference images (up to 10).
2Write a prompt describing the edit or composition.
3Optionally set makeup, lighting, shadows, and pose mode.
4Generate — Nano Banana Pro applies your prompt to the images.

Tips

Order images intentionally — first references often weigh more.
State clearly what to keep vs change in the prompt.

Open in Playground

Studios

StudiosNano Banana Pro

Fashion E-commerce Studio

Dress an Avatar in garments on a white / studio preset background.

Best for: Clean e-commerce catalogue shots (New Diffusion avatars).
Inputs: Avatar; pose (optional); garments one-by-one; background preset.
Outputs: Model wearing your garments on seamless studio BG.

How to use

1Select an Avatar from the library.
2Upload pose (optional) and Garment #1, then more garments one by one.
3Pick a white / studio background preset from the list.
4Generate — Nano Banana Pro dresses the avatar for e-commerce.

Tips

Use flat-lay or clear garment photos.
Pose library filters by avatar gender when available.

Open in Playground

StudiosNano Banana Pro

Fashion Creative Studio

Avatar + garments + your scenery photo for campaign looks.

Best for: Lookbook / editorial shots with a real location or mood scene.
Inputs: Avatar; garments first (then pose + scenery, any order).
Outputs: Dressed avatar standing in your scenery.

How to use

1Choose an avatar from the library (identity only).
2Upload garment(s), then a scenery photo.
3Describe pose and body framing in Extra details (e.g. half body, elegant standing) — there is no pose upload.
4Generate — the model is placed naturally in the scenery (Diversa-style).

Tips

Build the full outfit first, then pose and scenery — or any order.
Extra details add to the default campaign prompt; they do not replace it.

Open in Playground

StudiosNano Banana Pro → Z-Image Twin face refine

Twin Model E-commerce Studio

Twin LoRA identity + garments on a white / preset studio background.

Best for: Agency Twin catalogue shots with locked face.
Inputs: Twin Model; pose; garments; studio background preset.
Outputs: Try-on then Twin face-refined final.

How to use

1Select a Twin Model from the library.
2Upload pose and garments (flat lay or separate pieces).
3Choose a studio background preset (white / seamless).
4Generate — Nano Banana try-on, then Z-Image Twin face refine.

Tips

Always select the Twin before generating.
Prefer seamless white BG presets for e-commerce.

Open in Playground

StudiosNano Banana Pro → Z-Image Twin face refine

Twin Model Creative Studio

Twin + garments + uploaded scenery for creative campaigns.

Best for: Location / lifestyle Twin shoots with face refine.
Inputs: Twin Model; garments first (then pose + scenery, any order); optional addons.
Outputs: Scene try-on + Twin face-refined final.

How to use

1Select a Twin Model from the library.
2Add garments, then a scenery photo.
3Describe pose and body framing in Extra details — there is no pose upload.
4Generate — Twin wears your clothes naturally in the scenery.

Tips

Build the full outfit first, then pose and scenery — or any order.
Extra details add to the default campaign prompt; they do not replace it.

Open in Playground

StudiosNano Banana Pro (+ RMBG for transparent)

Packshot Studio

Turn a garment-on-person photo into Flatlay or Ghost Mannequin packshots.

Best for: E-commerce product packshots from worn-garment photos.
Inputs: 1 garment-on-person image; style Flatlay | Ghost Mannequin; white | transparent BG.
Outputs: Clean garment-only packshot.

How to use

1Upload one photo of the garment worn on a person.
2Choose Flatlay or Ghost Mannequin style.
3Choose White or Transparent background.
4Generate — Nano Banana Pro outputs a clean e-commerce packshot.

Tips

Use Frederic’s packshot prompts (built in): keep color, details, no folds, symmetric.
Transparent = white packshot then background removal for true alpha.

Open in Playground

StudiosNano Banana Pro

Lookbook Generator

Multiple campaign poses from one outfit setup.

Best for: Batch lookbook frames without re-uploading everything each time.
Inputs: Avatar / model refs + outfit; batch count; aspect; resolution.
Outputs: Several lookbook stills per run.

How to use

1Select an Avatar (or model reference) and upload the outfit.
2Set how many lookbook frames you want (batch).
3Choose aspect ratio and resolution.
4Generate — multiple campaign poses from one outfit.

Tips

If a high resolution fails, retry at 1K then upscale.
Keep garment references clean and well lit.

Open in Playground

Image Tools

Image ToolsComfyUI / SeedVR upscale

Resize/Upscale

Enlarge and sharpen existing images for delivery.

Best for: Taking approved generations to higher resolution for web or print.
Inputs: Source image; target resolution.
Outputs: Upscaled still.

How to use

1Upload the image you want to enlarge / sharpen.
2Choose the target resolution (up to the current max).
3Generate — Upscale rebuilds detail for a cleaner large output.
4Use results from Generation or the center preview.

Tips

Start from a clean, uncompressed source when possible.
Very large outputs may take longer on cold GPU starts.

Open in Playground

Image ToolsSmart overlay (no Comfy)

Fix Logo

Erase a wrong logo and place your exact logo PNG on the garment.

Best for: Brand-accurate logo fixes on fashion photos.
Inputs: Fashion photo + correct logo (transparent PNG best); drag position on preview.
Outputs: Photo with corrected logo.

How to use

1Upload the fashion photo (wrong logo OK).
2Upload the correct logo (transparent PNG best).
3Drag the logo on the preview over the old mark.
4Generate — old logo is erased, then your exact logo is placed.

Tips

Use a transparent PNG logo.
Align carefully on the preview before Generate.

Open in Playground

Image ToolsRMBG (Comfy)

Remove Background

Cut out subject to transparent or solid color background.

Best for: Packshots, composites, and clean product cutouts.
Inputs: Any photo; Alpha or color BG; batch 1–6.
Outputs: Cutout images.

How to use

1Upload a photo (person, product, or garment).
2Pick transparent (Alpha) or a solid color background.
3Choose how many images to generate (1–6).
4Generate — result shows in center + Generation.

Tips

Alpha mode for true transparency.
Fine hair / sheer fabrics may need a second pass.

Open in Playground

Image ToolsZ-Image Twin face refine

Refine Face

Apply Twin LoRA face refinement to an existing fashion photo.

Best for: Sharpening / locking Twin identity on a finished shot.
Inputs: Source photo + Twin Model; strength.
Outputs: Face-refined still.

How to use

1Upload a fashion / portrait photo to refine.
2Select your Twin model (required).
3Adjust strength if needed, then Generate.
4Result shows in center preview + Generation.

Tips

Select the correct Twin for the subject.
Use moderate strength to avoid over-smoothing.

Open in Playground

Image ToolsComfy mask + prompt inpaint

Change hairstyle

Paint a mask and restyle hair (or other areas) with a prompt.

Best for: Hair color / cut changes without rebuilding the whole image.
Inputs: Photo; painted mask; prompt; denoise ~0.85 for full restyle.
Outputs: Edited still.

How to use

1Upload a fashion / portrait photo.
2Paint a mask over the hair (or any area) to change.
3Write a clear prompt, e.g. short blonde bob, soft waves.
4Use denoise ~0.85 for a full restyle. Generate.

Tips

Mask only the area you want changed.
Clear prompts work best: “short blonde bob, soft waves”.

Open in Playground

Image ToolsNano Banana Pro

360° view generator

Generate multiple orbit angles around one product or subject.

Best for: Multi-angle product views from a single photo.
Inputs: One image; start angle / tilt / zoom; batch up to 6.
Outputs: Set of angled views.

How to use

1Upload one product or subject photo.
2Set start angle, tilt, and zoom if you want.
3Choose how many orbit views to generate (up to 6).
4Generate — Nano Banana Pro creates multi-angle views.

Tips

Centered subjects with simple backgrounds work best.
Use batch 4–6 for a full orbit set.

Open in Playground

Image ToolsNano Banana Pro

Style transfer

Avatar + garment + pose + style/vibe references → styled fashion image.

Best for: Full style builds from core references (Avatars + Poses libraries).
Inputs: Avatar → Garment #1 → Pose Reference → Style / vibe → optional addons.
Outputs: Styled fashion still.

How to use

1Upload Avatar, then Garment #1, Pose Reference, and Style / vibe.
2Optional garments open one by one after the core four.
3Add a short scene prompt if you want extra direction.
4Generate — Style transfer builds the look from your references.

Tips

Libraries: Face = Avatars, Pose = poses (not Twins / garments).
Upload progressively — next slot appears after each image.

Open in Playground

Image ToolsComfy depth + Twin LoRA

Twin Model Depth Pose Transfer

Transfer structure / pose to Twin identity from a reference photo.

Best for: Keeping pose structure while applying a Twin face.
Inputs: Reference photo (prefer white BG) + Twin Model + prompt.
Outputs: Pose-aware Twin image.

How to use

1Upload a fashion / pose reference photo.
2Tip: always use a white background image here to get a good output.
3Select your Twin model (required).
4Write a prompt, then Generate - result appears in center + Generation.

Tips

Always use a white background reference image for best results.
Neutral clothing on the reference helps avoid odd garment bleed.

Open in Playground

Image ToolsComfy DWPose + Twin LoRA

Twin Model Pose Transfer

Copy pose skeleton and apply Twin identity.

Best for: Matching a pose reference with your Twin model.
Inputs: Pose photo + Twin Model + prompt.
Outputs: Pose-matched Twin image.

How to use

1Upload a photo with the pose you want.
2Select your Twin model (required).
3Write a prompt describing the look.
4Generate - result appears in center + Generation.

Tips

Clear full-body pose photos work best.
Prefer neutral studio settings when possible.

Open in Playground

Image ToolsChat assistant

Prompt Builder

Chat to craft fashion prompts (and optionally generate from them).

Best for: Writers and teams who want structured, reusable prompts.
Inputs: Chat description; optional reference images.
Outputs: Ready prompts (and optional generated images).

How to use

1Describe what you want (subject, setting, style).
2Use the chat / builder to refine the prompt.
3Copy or send the finished prompt into an image workflow.
4Keep prompts specific: lighting, camera, fabric, mood.

Tips

Ask for lighting, lens, and fabric details.
Save strong prompts for reuse across studios.

Open in Playground

Video

VideoImage / text to video

Video

Animate a still or generate a short clip from a prompt.

Best for: Short fashion motion clips from approved stills.
Inputs: Reference image (mind max size) + motion prompt; or text mode.
Outputs: Video clip in Generation (download when ready).

How to use

1Upload a reference image (check max file size if noted).
2Write a short motion / scene prompt for image-to-video.
3Generate — wait for the clip to finish in Generation.
4Download from Generation when the preview is ready.

Tips

Keep source images under the stated size limit to avoid timeouts.
Image-to-video is more reliable than pure text-to-video today.

Open in Playground

FAQ

Where do I find each tool?+

Open Playground (Workflows). The left menu is grouped: Create, Edit, Studios, Image Tools, Video — same order as this documentation.

What is Avatar vs Twin Model?+

Avatars are face/body references for Fashion studios (no LoRA). Twin Models are trained LoRAs with face refine for Twin studios and Twin image tools.

Why is generation slow the first time?+

The GPU may wake from idle (cold start). Later runs in the same session are usually faster.

Can I use outputs commercially?+

Yes — images you generate are yours to use commercially under the Terms of Service.

Do failed generations cost credits?+

No. Credits are charged on successful runs. Failed jobs are refunded automatically.

Troubleshooting

Problem	What to try
No images generated / timeout	Retry once; if the GPU was cold, the second run often succeeds. Try 1K resolution on heavy studios, then upscale.
Depth Pose Transfer looks wrong	Use a white-background reference photo and a clear Twin selection. Avoid busy backgrounds on the pose source.
Fashion Creative looks “glued” on scenery	Use a high-quality scenery plate and the built-in campaign prompt style (clothes on model, standing in the setting).
Packshot still shows the person	Confirm Packshot Studio style (Flatlay / Ghost Mannequin). Re-run with a clear garment-on-person photo; use Transparent for cut-out.
Video timeout on upload	Reduce image file size (stay under the stated max MB) and prefer image-to-video mode.

Getting started

New Diffusion is a fashion AI platform. Open Playground (Workflows) to run tools. The left menu matches the groups below: Create, Edit, Studios, Image Tools, and Video.

1Sign in and open Playground from the top navigation.
2Pick a tool from the left menu — each card includes a How to use panel.
3Upload references, set options, Generate — results appear in preview and Generation.

Group	Tools	Jump to
Create	Create image, Create image with Twin model, Create Avatar
Edit	Edit Image
Studios	Fashion E-commerce Studio, Fashion Creative Studio, Twin Model E-commerce Studio, Twin Model Creative Studio, Packshot Studio, Lookbook Generator
Image Tools	Resize/Upscale, Fix Logo, Remove Background, Refine Face, Change hairstyle, 360° view generator, Style transfer, Twin Model Depth Pose Transfer, Twin Model Pose Transfer, Prompt Builder
Video	Video

All workflows

Full reference for every live tool — same names as the Playground menu. Use the table of contents to jump quickly.

Create

CreateZ-Image Turbo

Create image

Basic text-to-image. Type a prompt and generate — no Twin model or LoRA required.

Best for: Moodboards, concepts, backgrounds, and quick fashion ideas from text only.
Inputs: Text prompt; aspect ratio; megapixel; batch 1–6.
Outputs: Still images in Generation and Gallery.

How to use

1Write a clear text prompt describing the image you want.
2Choose aspect ratio (and megapixel if needed).
3Pick how many images to generate (1–6).
4Generate — Z-Image Turbo creates the image (no Twin / LoRA).

Tips

Be specific: subject, fabric, lighting, camera, mood.
Start at 2:3 or 1:1 for catalogue-style frames.

Open in Playground

CreateZ-Image Turbo + Twin LoRA

Create image with Twin model

Text-to-image locked to a Twin Model face from the library.

Best for: Consistent identity shots of a trained Twin without garments or scenery uploads.
Inputs: Prompt + Twin Model (required); strength; aspect; batch.
Outputs: Images with Twin identity applied.

How to use

1Write a prompt describing the look you want.
2Open Select Twin Model and pick one Twin face.
3Adjust model strength if needed, then set aspect ratio.
4Generate — result appears in center preview and Generation.

Tips

Include the Twin in Select Twin Model before generating.
Keep strength around 0.7–0.9 unless you need a softer likeness.

Open in Playground

CreateNano Banana Pro

Create Avatar

Generate clean portrait references for the Avatar Library.

Best for: Building new avatar faces (name, ref, gender, age) for Fashion studios.
Inputs: Description; gender; age range; aspect; batch.
Outputs: Portrait images; Super Admin can save to Avatar Library.

How to use

1Describe the avatar face you want (features, hair, age feel).
2Set gender and age range.
3Choose aspect ratio and how many images to generate.
4Generate — Super Admin can save the best result to Avatar Library.

Tips

Ask for front-facing head-and-shoulders, soft studio light, plain background.
Avoid props or hands covering the face.

Open in Playground

Edit

EditNano Banana Pro

Edit Image

Multi-image composition / text+image editing (up to 10 references). No Twin model.

Best for: Combining references, guided edits, and creative composites.
Inputs: 1–10 images + prompt; optional makeup / lighting / shadows / pose mode.
Outputs: Edited or composed stills.

How to use

1Upload one or more reference images (up to 10).
2Write a prompt describing the edit or composition.
3Optionally set makeup, lighting, shadows, and pose mode.
4Generate — Nano Banana Pro applies your prompt to the images.

Tips

Order images intentionally — first references often weigh more.
State clearly what to keep vs change in the prompt.

Open in Playground

Studios

StudiosNano Banana Pro

Fashion E-commerce Studio

Dress an Avatar in garments on a white / studio preset background.

Best for: Clean e-commerce catalogue shots (New Diffusion avatars).
Inputs: Avatar; pose (optional); garments one-by-one; background preset.
Outputs: Model wearing your garments on seamless studio BG.

How to use

1Select an Avatar from the library.
2Upload pose (optional) and Garment #1, then more garments one by one.
3Pick a white / studio background preset from the list.
4Generate — Nano Banana Pro dresses the avatar for e-commerce.

Tips

Use flat-lay or clear garment photos.
Pose library filters by avatar gender when available.

Open in Playground

StudiosNano Banana Pro

Fashion Creative Studio

Avatar + garments + your scenery photo for campaign looks.

Best for: Lookbook / editorial shots with a real location or mood scene.
Inputs: Avatar; garments first (then pose + scenery, any order).
Outputs: Dressed avatar standing in your scenery.

How to use

1Choose an avatar from the library (identity only).
2Upload garment(s), then a scenery photo.
3Describe pose and body framing in Extra details (e.g. half body, elegant standing) — there is no pose upload.
4Generate — the model is placed naturally in the scenery (Diversa-style).

Tips

Build the full outfit first, then pose and scenery — or any order.
Extra details add to the default campaign prompt; they do not replace it.

Open in Playground

StudiosNano Banana Pro → Z-Image Twin face refine

Twin Model E-commerce Studio

Twin LoRA identity + garments on a white / preset studio background.

Best for: Agency Twin catalogue shots with locked face.
Inputs: Twin Model; pose; garments; studio background preset.
Outputs: Try-on then Twin face-refined final.

How to use

1Select a Twin Model from the library.
2Upload pose and garments (flat lay or separate pieces).
3Choose a studio background preset (white / seamless).
4Generate — Nano Banana try-on, then Z-Image Twin face refine.

Tips

Always select the Twin before generating.
Prefer seamless white BG presets for e-commerce.

Open in Playground

StudiosNano Banana Pro → Z-Image Twin face refine

Twin Model Creative Studio

Twin + garments + uploaded scenery for creative campaigns.

Best for: Location / lifestyle Twin shoots with face refine.
Inputs: Twin Model; garments first (then pose + scenery, any order); optional addons.
Outputs: Scene try-on + Twin face-refined final.

How to use

1Select a Twin Model from the library.
2Add garments, then a scenery photo.
3Describe pose and body framing in Extra details — there is no pose upload.
4Generate — Twin wears your clothes naturally in the scenery.

Tips

Build the full outfit first, then pose and scenery — or any order.
Extra details add to the default campaign prompt; they do not replace it.

Open in Playground

StudiosNano Banana Pro (+ RMBG for transparent)

Packshot Studio

Turn a garment-on-person photo into Flatlay or Ghost Mannequin packshots.

Best for: E-commerce product packshots from worn-garment photos.
Inputs: 1 garment-on-person image; style Flatlay | Ghost Mannequin; white | transparent BG.
Outputs: Clean garment-only packshot.

How to use

1Upload one photo of the garment worn on a person.
2Choose Flatlay or Ghost Mannequin style.
3Choose White or Transparent background.
4Generate — Nano Banana Pro outputs a clean e-commerce packshot.

Tips

Use Frederic’s packshot prompts (built in): keep color, details, no folds, symmetric.
Transparent = white packshot then background removal for true alpha.

Open in Playground

StudiosNano Banana Pro

Lookbook Generator

Multiple campaign poses from one outfit setup.

Best for: Batch lookbook frames without re-uploading everything each time.
Inputs: Avatar / model refs + outfit; batch count; aspect; resolution.
Outputs: Several lookbook stills per run.

How to use

1Select an Avatar (or model reference) and upload the outfit.
2Set how many lookbook frames you want (batch).
3Choose aspect ratio and resolution.
4Generate — multiple campaign poses from one outfit.

Tips

If a high resolution fails, retry at 1K then upscale.
Keep garment references clean and well lit.

Open in Playground

Image Tools

Image ToolsComfyUI / SeedVR upscale

Resize/Upscale

Enlarge and sharpen existing images for delivery.

Best for: Taking approved generations to higher resolution for web or print.
Inputs: Source image; target resolution.
Outputs: Upscaled still.

How to use

1Upload the image you want to enlarge / sharpen.
2Choose the target resolution (up to the current max).
3Generate — Upscale rebuilds detail for a cleaner large output.
4Use results from Generation or the center preview.

Tips

Start from a clean, uncompressed source when possible.
Very large outputs may take longer on cold GPU starts.

Open in Playground

Image ToolsSmart overlay (no Comfy)

Fix Logo

Erase a wrong logo and place your exact logo PNG on the garment.

Best for: Brand-accurate logo fixes on fashion photos.
Inputs: Fashion photo + correct logo (transparent PNG best); drag position on preview.
Outputs: Photo with corrected logo.

How to use

1Upload the fashion photo (wrong logo OK).
2Upload the correct logo (transparent PNG best).
3Drag the logo on the preview over the old mark.
4Generate — old logo is erased, then your exact logo is placed.

Tips

Use a transparent PNG logo.
Align carefully on the preview before Generate.

Open in Playground

Image ToolsRMBG (Comfy)

Remove Background

Cut out subject to transparent or solid color background.

Best for: Packshots, composites, and clean product cutouts.
Inputs: Any photo; Alpha or color BG; batch 1–6.
Outputs: Cutout images.

How to use

1Upload a photo (person, product, or garment).
2Pick transparent (Alpha) or a solid color background.
3Choose how many images to generate (1–6).
4Generate — result shows in center + Generation.

Tips

Alpha mode for true transparency.
Fine hair / sheer fabrics may need a second pass.

Open in Playground

Image ToolsZ-Image Twin face refine

Refine Face

Apply Twin LoRA face refinement to an existing fashion photo.

Best for: Sharpening / locking Twin identity on a finished shot.
Inputs: Source photo + Twin Model; strength.
Outputs: Face-refined still.

How to use

1Upload a fashion / portrait photo to refine.
2Select your Twin model (required).
3Adjust strength if needed, then Generate.
4Result shows in center preview + Generation.

Tips

Select the correct Twin for the subject.
Use moderate strength to avoid over-smoothing.

Open in Playground

Image ToolsComfy mask + prompt inpaint

Change hairstyle

Paint a mask and restyle hair (or other areas) with a prompt.

Best for: Hair color / cut changes without rebuilding the whole image.
Inputs: Photo; painted mask; prompt; denoise ~0.85 for full restyle.
Outputs: Edited still.

How to use

1Upload a fashion / portrait photo.
2Paint a mask over the hair (or any area) to change.
3Write a clear prompt, e.g. short blonde bob, soft waves.
4Use denoise ~0.85 for a full restyle. Generate.

Tips

Mask only the area you want changed.
Clear prompts work best: “short blonde bob, soft waves”.

Open in Playground

Image ToolsNano Banana Pro

360° view generator

Generate multiple orbit angles around one product or subject.

Best for: Multi-angle product views from a single photo.
Inputs: One image; start angle / tilt / zoom; batch up to 6.
Outputs: Set of angled views.

How to use

1Upload one product or subject photo.
2Set start angle, tilt, and zoom if you want.
3Choose how many orbit views to generate (up to 6).
4Generate — Nano Banana Pro creates multi-angle views.

Tips

Centered subjects with simple backgrounds work best.
Use batch 4–6 for a full orbit set.

Open in Playground

Image ToolsNano Banana Pro

Style transfer

Avatar + garment + pose + style/vibe references → styled fashion image.

Best for: Full style builds from core references (Avatars + Poses libraries).
Inputs: Avatar → Garment #1 → Pose Reference → Style / vibe → optional addons.
Outputs: Styled fashion still.

How to use

1Upload Avatar, then Garment #1, Pose Reference, and Style / vibe.
2Optional garments open one by one after the core four.
3Add a short scene prompt if you want extra direction.
4Generate — Style transfer builds the look from your references.

Tips

Libraries: Face = Avatars, Pose = poses (not Twins / garments).
Upload progressively — next slot appears after each image.

Open in Playground

Image ToolsComfy depth + Twin LoRA

Twin Model Depth Pose Transfer

Transfer structure / pose to Twin identity from a reference photo.

Best for: Keeping pose structure while applying a Twin face.
Inputs: Reference photo (prefer white BG) + Twin Model + prompt.
Outputs: Pose-aware Twin image.

How to use

1Upload a fashion / pose reference photo.
2Tip: always use a white background image here to get a good output.
3Select your Twin model (required).
4Write a prompt, then Generate - result appears in center + Generation.

Tips

Always use a white background reference image for best results.
Neutral clothing on the reference helps avoid odd garment bleed.

Open in Playground

Image ToolsComfy DWPose + Twin LoRA

Twin Model Pose Transfer

Copy pose skeleton and apply Twin identity.

Best for: Matching a pose reference with your Twin model.
Inputs: Pose photo + Twin Model + prompt.
Outputs: Pose-matched Twin image.

How to use

1Upload a photo with the pose you want.
2Select your Twin model (required).
3Write a prompt describing the look.
4Generate - result appears in center + Generation.

Tips

Clear full-body pose photos work best.
Prefer neutral studio settings when possible.

Open in Playground

Image ToolsChat assistant

Prompt Builder

Chat to craft fashion prompts (and optionally generate from them).

Best for: Writers and teams who want structured, reusable prompts.
Inputs: Chat description; optional reference images.
Outputs: Ready prompts (and optional generated images).

How to use

1Describe what you want (subject, setting, style).
2Use the chat / builder to refine the prompt.
3Copy or send the finished prompt into an image workflow.
4Keep prompts specific: lighting, camera, fabric, mood.

Tips

Ask for lighting, lens, and fabric details.
Save strong prompts for reuse across studios.

Open in Playground

Video

VideoImage / text to video

Video

Animate a still or generate a short clip from a prompt.

Best for: Short fashion motion clips from approved stills.
Inputs: Reference image (mind max size) + motion prompt; or text mode.
Outputs: Video clip in Generation (download when ready).

How to use

1Upload a reference image (check max file size if noted).
2Write a short motion / scene prompt for image-to-video.
3Generate — wait for the clip to finish in Generation.
4Download from Generation when the preview is ready.

Tips

Keep source images under the stated size limit to avoid timeouts.
Image-to-video is more reliable than pure text-to-video today.

Open in Playground

FAQ

Where do I find each tool?+

Open Playground (Workflows). The left menu is grouped: Create, Edit, Studios, Image Tools, Video — same order as this documentation.

What is Avatar vs Twin Model?+

Avatars are face/body references for Fashion studios (no LoRA). Twin Models are trained LoRAs with face refine for Twin studios and Twin image tools.

Why is generation slow the first time?+

The GPU may wake from idle (cold start). Later runs in the same session are usually faster.

Can I use outputs commercially?+

Yes — images you generate are yours to use commercially under the Terms of Service.

Do failed generations cost credits?+

No. Credits are charged on successful runs. Failed jobs are refunded automatically.

Troubleshooting

Problem	What to try
No images generated / timeout	Retry once; if the GPU was cold, the second run often succeeds. Try 1K resolution on heavy studios, then upscale.
Depth Pose Transfer looks wrong	Use a white-background reference photo and a clear Twin selection. Avoid busy backgrounds on the pose source.
Fashion Creative looks “glued” on scenery	Use a high-quality scenery plate and the built-in campaign prompt style (clothes on model, standing in the setting).
Packshot still shows the person	Confirm Packshot Studio style (Flatlay / Ghost Mannequin). Re-run with a clear garment-on-person photo; use Transparent for cut-out.
Video timeout on upload	Reduce image file size (stay under the stated max MB) and prefer image-to-video mode.