# Holala Academy
Source: https://docs.holala.ai/academy
Short lessons and walkthrough videos that teach the holala workflow end to end.
Holala Academy is the video library for holala. Every lesson is a short, real walkthrough of one template — what to upload, what to pick, and what comes out. Watch the one that matches your job and copy the flow.
New to holala? Start with the first lesson below — it covers the core flow (upload → template → Generate) that every other template reuses.
## Start here
### Product Replacement — create on-model images without a studio
Put your product on a model without booking a shoot.
## Scenes & backgrounds
### Background Replacement — move your product into a new scene
### White Background — marketplace-ready packshots
### Style Duplicator — apply a look you love to your own product
## Models & faces
### Model Swapper — change the model, keep the product
### Model Remover — strip the model out, keep a clean product image
### Face Generator — create a consistent AI model face
## Garment presentation
### Ghost Mannequin Creator — hollow-form shots without a mannequin
### On-Model to Flat-lay — turn a model shot into a flat-lay
### Closeup Texture Generator — macro fabric shots that sell the material
## Special categories & quality
### Kids Image Generator — on-model imagery for kids collections
### Lingerie & Swimwear Generator — brand-safe on-model imagery
### Quality Enhancer — sharpen and upscale low-resolution imagery
## Read the matching guides
Every template's written guide — inputs, steps, and tips for the workflow you just watched.
The power-user habits that save credits and keep a whole collection consistent.
# Common Mistakes
Source: https://docs.holala.ai/common-mistakes
The input mistakes that quietly ruin holala results, and how to fix each one before you generate.
Most disappointing results trace back to one thing: the input photo, not the AI. holala refines what your photo already shows — it does not invent angles, details, or context that aren't there. Fix the input, and the output follows.
Here are the mistakes that come up most often, and the fix for each.
You uploaded a three-quarter shot but expected a straight-on, front-facing result. holala can restyle a scene, but it cannot rotate a product to an angle your photo never captured.
**Fix:** Match your input angle to your desired output angle. If you need a front-facing marketing image, shoot the product front-facing. See [Input & Uploads](/input-uploads) for angle guidance per template.
A cluttered backdrop, a patterned surface, or a busy room behind the product competes with the AI's read of the product itself, and often bleeds into the generated scene in odd ways.
**Fix:** Shoot on a plain, neutral background whenever possible. The simpler your backdrop, the cleaner holala's read of your product edges, color, and shape.
If the source image is small, blurry, or cropped so part of the product is cut off, holala has less real detail to work with and has to guess at what's missing.
**Fix:** Upload the highest-resolution photo you have, fully framed with the whole product visible. Check [Input & Uploads](/input-uploads) for minimum recommendations before you generate.
Uploading a photo with two or more products together confuses which item the template should feature, and Templates are built as one-job functions for a single product.
**Fix:** Upload one product per generation. Run separate jobs for separate products, then organize the results in your Library.
1K is a quick test resolution. Publishing a 1K result as your final marketing image gives you softer detail than the platform is capable of.
**Fix:** Test your template and inputs at 1K to confirm the composition works, then re-run at 2K or 4K for anything you'll actually publish. See [Credits](/credits) for how resolution affects credit cost.
An optional reference image is meant to guide style, scene, or lighting. A reference from the wrong product category, a busy or cluttered reference, or a low-resolution reference pulls the generation in a conflicting direction instead of helping it.
**Fix:** Choose a reference that matches your product's category and is itself a clean, high-resolution image. If a reference is fighting your results, remove it and re-test with the product image alone.
holala refines what's already in your photo — it does not fabricate a side, texture, or feature your input never showed. If your input doesn't show it, the output won't either.
**Fix:** Before you generate, ask: does this input already contain everything I want the final image to show? If not, reshoot the input rather than expecting the generation to fill the gap.
Every one of these mistakes traces back to the input photo. When a result looks off, check your source image first before assuming the template or settings are wrong.
## Quick Do / Don't reference
**Do**
* Match input angle to desired output angle
* Use a plain, neutral background
* Upload full-resolution, fully-framed photos
* One product per generation
* Test at 1K, finalize at 2K or 4K
* Pick a reference in the same category, clean and high-res
**Don't**
* Expect a front result from a side-angle input
* Shoot on busy or patterned backdrops
* Upload cropped or blurry source images
* Combine multiple products in one frame
* Publish 1K as your final image
* Use a low-quality or mismatched reference image
## Related reading
The full guide to preparing product and reference images before you generate.
How resolution and variation count affect what a generation costs.
# Credits
Source: https://docs.holala.ai/credits
Credits are how you pay for generations in holala. Every image the AI creates costs credits. No subscriptions to decode, no per-prompt haggling — you spend credits, you get images.
## One image, one charge
A credit is spent for **each image you generate**. Ask for one image, you're charged once. Ask for four variations, you're charged for four.
Two things drive how much a single generation costs:
Higher resolution means a bigger, sharper file — and a higher credit cost per image. **1K** is the cheapest, **2K** costs more, **4K** costs the most.
Each variation is its own image, so each one is charged separately. Four variations cost four times a single image at the same resolution.
Put simply: **cost = resolution per image × how many images you ask for.** Turn either dial up and the generation costs more.
**1K is for tests only.** Use it to check framing, model choice, or scene direction before committing. The quality usually isn't enough for a final. Generate your finished images at **2K or 4K**.
## Trial credits to start
New demo accounts get **trial credits added automatically** — no purchase, no card. That's enough to run real generations, try a few templates, and see the output quality for yourself before you decide.
Spend your trial credits on things you actually want to keep. Run a quick 1K test to lock in your setup, then generate the final at 2K or 4K.
## Watching your balance
Your credit balance is always visible in the STUDIO while you work. Check it before a big batch — especially before a bulk generation or a run of high-resolution variations, since those add up fastest.
## Save credits
A few habits stretch your balance a long way:
* **Test at 1K first.** Dial in the template, model, and scene at the cheapest resolution, then re-run once at 2K or 4K for the keeper.
* **Prep your inputs well.** A clean product photo — plain background, single product, correct angle, sharp and uncropped — gets a good result the first time and saves you from re-rolls. See [Input uploads](/input-uploads) and [Common mistakes](/mistakes).
* **Generate fewer variations once you're dialed in.** Explore with a couple of options, then when the look is right, produce just the finals you need instead of a large batch.
Every generation spends credits, even ones you end up not using. Getting your product image and reference right before you click **Generate** is the single best way to avoid wasted credits.
## Need more credits?
Trial credits are meant to get you started. When you're ready to keep going, sign up for a full holala account.
Request access. Once you're accepted, you'll get the sign-up link.
Head to STUDIO, check your balance, and start generating.
# Generation Steps
Source: https://docs.holala.ai/development
A generation in Holala is completed in five steps. Follow them in order.
## Step-by-step generation
Select a template based on what you want to produce — product placement, model swap, ghost mannequin, etc. Each template may require different inputs.
Drag and drop your prepared product image. If My Store is connected, you can select existing products from the list.
Add a reference image for style guidance. You can choose from the library or upload a new image.
For model swap or ghost mannequin templates, choose a model from the library. Filter by gender, ethnicity, and body part.
Select 1K, 2K, or 4K resolution. Set the number of variations. Each additional variation consumes credits.
Click Generate. When complete, results are automatically saved to your Library. Download individually or in bulk.
**Credit note:** 1K uses the fewest credits, 4K the most. Use 1K for test runs, 2K or 4K for final output. 1K quality is generally not sufficient for publishing.
## What can you do with the result?
Export as JPG, PNG, or WebP.
Push the image directly to your connected store — Shopify, IKAS, etc.
Generations are saved automatically. Return to them any time.
## Key concepts
Defines the generation purpose — product placement, ghost mannequin, and more.
Everything you need to generate ultra-high-quality content inside holala.
Each generation consumes credits. 1K, 2K, and 4K resolutions use different amounts.
The combination of product image, reference image, model selection, and prompts.
# Editor
Source: https://docs.holala.ai/editor
The Editor is where you refine a generated result and get it export-ready — review variations, polish the winner, and download in the format you need.
The **Editor** is where a generated image becomes a finished one. After Studio creates your variations, the Editor is where you review them, pick the strongest, and get it export-ready.
holala is in **alfa (V.1.0)** and the Editor is growing. This page covers what it's for; expect more refinement tools over time.
## What the Editor is for
Line up the versions from a generation side by side and choose the best frame instead of settling for the first.
Take a result further — send it through a template like Quality Enhancer to sharpen and upscale before export.
Download the final image as JPG, PNG, or WebP, at the resolution you generated.
If you've connected My Store (Shopify, IKAS), push the finished image straight back to your product listing.
## A typical refine pass
Every generation is saved automatically. Open the one you want to finish.
Review the variations from the run and pick the strongest one.
Not sharp enough? Send it through [Quality Enhancer](/templates). Want a different angle or scene? Re-run it in [Studio](/studio) with the right template.
Download it, or push it back to your store.
The best editorial results are usually **chosen from a set**, not landed in one shot. Generate a few variations in [Studio](/studio) so the Editor has strong options to pick from. See [Editorial Looks](/editorial).
Generate more variations or try another template.
Push results from clean to campaign-grade.
# Editorial Looks
Source: https://docs.holala.ai/editorial
A clean product photo and a full editorial campaign use the same STUDIO. The difference is not luck — it is five deliberate choices. Set them well and the output reads like a brand shoot, not a catalog cutout.
No prompt-writing required. Editorial quality comes from your **inputs** — the reference you feed, the resolution you pick, and the scene you choose — not from clever wording.
## The five levers
A strong style/lighting reference is the single biggest lever. It hands the AI your mood, your palette, and your light in one image.
1K is for quick tests only. Campaign work needs 2K or 4K so texture, edges, and fabric hold up at full size.
A plain white sweep says "packshot." An intentional set — surface, prop, atmosphere — says "campaign."
The output angle follows the input. A hero front shot starts from a front-facing product photo. The AI can't invent a view your input doesn't show.
## The recipe
For a styled scene, reach for **Background Generation** or **Product Placement**. For a campaign on a person, use **Product-to-On-Model**. Each template does one job — see [all templates](/templates).
Plain background, single product, centered, sharp, full in frame, shot at the angle you want in the final. This is your foundation — a weak input caps the output.
Attach a reference that shows the exact lighting and color tone you want. Keep it to one clear atmosphere, same product category as yours. This is what pushes the result from clean to editorial.
Filter by gender, ethnicity, and body part to match the campaign's cast. Skip for still-life templates.
Higher resolution for final quality; multiple variations so you can pick the strongest frame instead of settling for the first.
Review the batch in your Library. Swap the reference, nudge the scene, and re-run. Editorial results are chosen from a set, rarely landed in one shot.
## What a reference controls
The reference image is where the editorial "feel" lives. A good one does the heavy lifting; a mismatched one fights your product.
One clear style and atmosphere\
Shows the lighting and color tone you want\
Same product category as yours\
Clean, high-resolution, uncropped
Busy scene with competing focal points\
A different product category\
Low resolution or heavily cropped\
Multiple products in one frame
No reference on hand? Open the **Inspiration** library and pull a curated reference scene. It is built for exactly this — proven lighting and mood you can drop straight into STUDIO.
## Before / after: framing the difference
The jump from packshot to editorial is a framing decision, not a filter.
Product on white, even light, no story. Correct, but forgettable — it does a job and nothing more.
![A single product centered on a plain white background with flat, even studio lighting and no surrounding scene]()
Same product, styled scene, directional light, a reference-driven palette. Now it reads like a campaign frame you'd run in an ad.
Resolution is not a finishing step you can bolt on later. Generate the editorial frame at **2K or 4K** from the start — 1K output usually isn't sharp enough for a campaign, and there's no upgrade after the fact. Export as JPG, PNG, or WebP.
## Dial it in
Treat the first batch as a draft, not a verdict. Small, single-variable changes move the look fastest.
Swap in a reference with stronger directional lighting. The reference drives the light more than any other input — change it before you change anything else.
Choose a reference whose palette already matches your brand. The AI pulls color atmosphere straight from the reference, so pick the mood you want to land.
Re-upload a product photo shot at the angle you want. The output can't show a front the input never captured — the angle is set at the input stage.
Request more variations and try a bolder reference scene. A wider set gives you a stronger frame to choose from.
Every generation auto-saves to your **Library**, so you can line up variations side by side and pick the winner — no re-running to recover a frame you liked.
## Keep going
The input and reference errors that quietly flatten editorial output — and how to sidestep each one.
Find the right function for the shot: scenes, on-model, ghost mannequin, closeup textile, and more.
# General Videos
Source: https://docs.holala.ai/general-videos
[Part 01 - General Overview](https://screen.studio/share/fhEqcGMH?private-access=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJ2ZXJzaW9uIjoiMi4wIiwic2hhcmVhYmxlTGlua0lkIjoiMTE4Y2E1MWItMzU2Ni00Nzg0LTg5ZTgtOTNmY2FhZDY5NGJlIiwiaWF0IjoxNzc1MjM4ODY3fQ.8WIL6xTAFmr_Cltk5enpMQEDwo3VFAcC_Oo3vud07ig)
# Getting Started
Source: https://docs.holala.ai/index
Welcome to Holala. This page is designed for first-time users setting up the platform.
Upload a raw product photo and get back professional, brand-consistent images — no studio, no photographer, no prompt-writing. You work inside **Studio**: pick a **Template** for the job, add your inputs, and click **Generate**. Results save straight to your Library. Video is coming soon.
## What you can do
Drop your product into a clean, styled scene. Guide the look with an optional reference image.
Put your product on a model, change the model, or remove the model entirely — filter by gender, ethnicity, and body part.
Turn a mannequin or flat-lay shot into a clean, hollow-body product image.
Replace or generate a new background — or drop to a crisp white background for catalog listings.
Run the same template across many products at once instead of one by one.
Quality Enhancer, View/Angle Generator, Jewellery, Closeup Textile, and more. [Browse all templates](/templates).
## From sign-up to your first image
Request access on the [Waiting List form](https://holala.ai/holala-waiting-list).
When you're accepted, complete the [Sign-up form](https://tally.so/r/A7DEOy). You'll get a demo account with trial credits added automatically.
Log in to the app at [app.holala.ai](https://app.holala.ai) and open Studio.
Choose the template that matches the job you want done. Each template does one thing. [See the full list](/templates).
The product image is required. Use a plain background, single centered product, correct angle, high resolution. [Prep tips →](/input-uploads)
Add a reference image to guide style, scene, and lighting. For model templates, select a model by gender, ethnicity, or body part.
Pick 1K, 2K, or 4K and how many variations to generate. Use 2K or 4K for final output; 1K is for quick tests only. [How credits work →](/credits)
Click Generate. Your results are auto-saved to your Library in JPG, PNG, or WebP.
New demo accounts get trial credits automatically — enough to try a few generations right away.
## Key concepts
Every template is a function that does one job. Start here to find the right one.
The product photo rules that make or break a result — plain background, right angle, full product in frame.
What each generation costs, and why 1K is for tests while 2K/4K are for final output.
The full Studio flow, inputs, Library, and connecting My Store (Shopify, IKAS).
Getting a front-facing result needs a front-facing input — the AI can't invent a view your photo doesn't show. See the [input tips](/input-uploads) before your first run.
## Need inspiration first?
Curated presets, model faces, and reference scenes to copy from and remix into your own generations.
# Product Images
Source: https://docs.holala.ai/input-uploads
The single biggest factor in your results is the **product photo you upload**. holala reads what's in the frame and rebuilds it into a professional, brand-consistent image. Feed it a clean, correctly-shot input and every template — from Product Placement to Ghost Mannequin — has more to work with.
Get this page right once and you'll stop re-generating.
This page is about the **product image** — the one required input on every generation. For the optional **reference image** that guides style, scene, and lighting, see [Reference Images](/mistakes).
## Why the input decides the output
holala doesn't imagine your product from scratch. It works **from the pixels you give it** — the shape, color, texture, and angle in your photo. It can relight a scene, change a background, place your product on a model, or clean up a mannequin. It can't recover detail that was never captured.
Think of it like this: a sharp, well-lit, front-facing photo gives the model an honest starting point. A blurry, cropped, or badly-angled photo forces it to guess — and guesses are where artifacts come from.
Good input in, professional output out. The five minutes you spend on a clean product shot save you credits you'd otherwise burn re-generating.
## Correct vs. Incorrect
Match your product photo to the left column. Every item on the right makes results worse.
**Do this**
* **Plain or white background** — nothing competing with the product
* **One product, centered** in the frame
* **Correct angle** — shoot the view you want as output
* **High resolution, sharp** — in focus, not grainy
* **Full product visible** — nothing cut off at the edges
* **Even, natural lighting** — true colors, no harsh glare
**Avoid this**
* **Busy or patterned background** the AI has to fight
* **Several products** in one frame
* **Wrong angle** for the result you're after
* **Blurry, low-res, or pixelated** source
* **Cropped product** — edges or corners cut off
* **Heavy shadows or blown-out glare** hiding detail
![Side-by-side comparison of a correct product input photo versus an incorrect one]()
## The angle rule
This is the mistake people make most, so it gets its own callout.
**A front-facing output needs a front-facing input.** holala can relight, restage, and reframe your product — but it **cannot invent a view that isn't in the photo**. Upload a side shot and you'll get a side result. If you need the product seen from the front, shoot it from the front.
The same logic applies to every angle. Want a three-quarter hero shot? Start from a three-quarter photo. Need the back of the product? Photograph the back. The model refines the angle you give it — it doesn't rotate the object in 3D.
Need multiple angles from a single photo? That's a specific job for the **View/Angle Generator** and **Side Generation** templates — see [Templates](/templates). For everything else, shoot the angle you want.
## Cropping and background cleanup
Two quick fixes that make a visible difference:
Leave a little breathing room on all four sides. A product touching or bleeding past the edge gives the AI an incomplete shape to work from, which shows up as cut-off or warped edges in the output.
A plain or white background is easiest for the model to isolate the product from. If your photo has a distracting surface, table, or scene, remove it first. The cleaner the separation between product and background, the crisper the result — especially for White Background, Ghost Mannequin, and Product Placement.
Upload a single product. Multiple items confuse which one to work on and split the model's attention. Generate them one at a time.
## Category-specific prep
The fundamentals above apply to everything. These notes cover what matters most **per product type**.
Lay the garment flat or shoot it on a clean form so the **cut and silhouette read clearly**. Smooth out wrinkles and folds — the AI treats creases as real, so they carry into the output. Make sure the full garment is in frame, including hems, cuffs, and collar. For a front-on catalog look, shoot front-on.
Shoe shape is defined by the profile. A clean **side or three-quarter view** captures the silhouette best. Keep both the sole line and the upper visible, and avoid steep top-down angles that flatten the shape. Pairs should be arranged consistently, but a single clean shoe usually gives the strongest result.
Frames are small and detail-dense, so **resolution matters more than usual**. Shoot straight-on to keep both lenses and temples symmetrical, on a plain background. Watch for reflections and glare on the lenses — they hide the frame shape and confuse the model. Diffuse, even light keeps the frame edges crisp.
Small, reflective, and all about sparkle — get **as close and as sharp as possible**. Fill the frame with the piece, use even lighting to control harsh glints, and keep the background plain so gemstones and metal stand out. Clean the item first; dust and smudges are very visible at this scale. See the dedicated **Jewellery Image Generation** template in [Templates](/templates).
**Labels must stay legible.** Shoot packaging straight-on and in high resolution so brand names, text, and logos are sharp and readable in the source. If the label is blurry, angled, or too small to read in your photo, it won't come out clean in the result — and product text is exactly what customers scrutinize.
Keep bottles, tubes, and jars upright and centered, on a plain background, with even light that avoids blown-out highlights on glossy or reflective packaging.
Fit and drape are everything, so shoot on a clean form or lay flat so the **shape and straps read clearly**. Keep the full piece in frame with even lighting that shows the true fabric color and texture. holala has purpose-built **Lingerie & Swimwear Generation** — see [Templates](/templates) — so start from a clean, well-lit product shot for the best on-model results.
Capture the **full drape and length** — modest garments are defined by how the fabric falls, so don't crop the flow. Lay flat or shoot on a form against a plain background, smooth out folds you don't want, and light evenly to show the true color and weave. Keep the whole garment in frame from top to hem.
Structure and texture sell leather. A **three-quarter angle** shows the shape, depth, and hardware best. Keep straps, handles, and buckles fully in frame and unobscured, and use even light so the grain and stitching stay visible without harsh glare on shiny finishes. Stuff soft bags lightly so they hold their intended shape.
## Next: the reference image
Your product photo is the required input. The **optional reference image** is how you steer the style, scene, and lighting of the result — and it has its own set of do's and don'ts.
How to pick a reference that guides mood, lighting, and scene — and the common mistakes to avoid.
Browse every template and see which input each one expects.
# Inspiration
Source: https://docs.holala.ai/inspiration
General Categories
[https://pin.it/3og7zgYBT](https://pin.it/3og7zgYBT)
[https://pin.it/36Z5wlPzo](https://pin.it/36Z5wlPzo)
[https://pin.it/2pU8p7ePD](https://pin.it/2pU8p7ePD)
[https://pin.it/IWThxgFx4](https://pin.it/IWThxgFx4)
[https://pin.it/dEaWPInG5](https://pin.it/dEaWPInG5)
[https://pin.it/3ENbUHC9N](https://pin.it/3ENbUHC9N)
[https://pin.it/2S3EebeyF](https://pin.it/2S3EebeyF)
[https://pin.it/4byA0viLi](https://pin.it/4byA0viLi)
***
"Faces
[https://pin.it/5WhZRkkOm](https://pin.it/5WhZRkkOm)
[https://pin.it/4ZMXvgwBw](https://pin.it/4ZMXvgwBw)
***
# Reference Images
Source: https://docs.holala.ai/mistakes
A reference image tells holala the *look* you want. Your product image says **what** to render; the reference says **how** it should feel — the scene, the pose, the lighting, the color tone. It's the fastest way to steer a result without writing a prompt.
A reference is **optional**. Most templates run fine on your product image alone. Add a reference only when you have a specific look in mind — a scene, a mood, a lighting style you want to match.
## What a reference controls
Think of the reference as a mood board, not a copy target. holala reads the atmosphere from it and applies that feel to *your* product. It does not paste your product into the reference photo, and it does not swap in the reference's product.
A good reference carries **one** clear idea: a single scene, a single lighting setup, a single color story. When you hand it one clean signal, you get a clean, brand-consistent result.
## Good vs. weak references
**A strong reference**
* One clear style or atmosphere
* Shows the exact lighting and color tone you want
* Same product category as yours
* Clean, uncluttered composition
* High resolution, sharp, full frame
**A weak reference**
* Busy scene with competing subjects
* A different product category than yours
* Mixed or contradictory lighting
* Low resolution or blurry
* Cropped, or several products in one frame
Category match matters more than people expect. A reference in the *same* category as your product (bag → bag, sneaker → sneaker) gives the model the right proportions and context to work from.
## Three types of reference
Different jobs call for different references. Pick the one that matches what you're trying to steer.
**Controls where the product sits** — the setting, surface, and props around it.
Use a scene reference when you want your product placed in a specific environment: a marble countertop, a sunlit café table, a soft studio sweep. holala borrows the *setting and mood*, then drops your product into a result that matches it.
**Controls how the model or product is posed** — angle, stance, framing.
Use a pose reference for model and on-model templates when you want a particular body position or camera angle. It guides the *posture and composition* of the shot, not the model's identity.
**Controls the light and color grade** — direction, softness, and tone.
Use a style/lighting reference to match a mood: warm golden-hour glow, hard editorial shadows, clean high-key white. It steers the *lighting and color treatment* of the final image.
**Common reference mistakes**
* **Busy scenes** — multiple subjects confuse the model about what to echo.
* **Category mismatch** — a furniture reference for a lipstick pulls the result off. Keep the category the same.
* **Contradictory lighting** — a reference lit three different ways gives no clear signal to follow.
* **Low res, cropped, or multi-product** — the same quality bar as your product photo applies here. Blurry or cropped references produce blurry, off-frame direction.
* **Expecting a copy-paste** — the reference guides atmosphere; it does not transplant your product into it, nor its product into your result.
## Next steps
The reference only helps if your product photo is clean. See the input prep rules for the required product image.
Every generation costs credits. More variations and higher resolution cost more — plan before you generate.
# Model Faces
Source: https://docs.holala.ai/model-faces
Pick a model for your model templates, keep the same face across a whole drop, and choose between generating a new face or swapping one onto an existing frame.
Model templates put your product on a person. Before you Generate, you choose **who** that person is — from the model library, filtered to match your product and your market. Pick once, reuse everywhere, and the same face carries across your whole collection.
Model selection only applies to **model templates** — Product-to-On-Model, Model Swap, Jewellery Image Generation, Kids Image Generation, Lingerie & Swimwear Generation, and the rest. Object-only templates like White Background or Product Placement don't ask for a model.
## Pick a model from the library
Open a model template in the STUDIO, upload your product, then choose your model. The model library is built in — no casting, no shoot. Narrow it down with three filters until the model fits your product and audience:
Filter to the gender you're merchandising for.
Match the market you're selling into.
For close-crop templates — a hand for jewellery, a wrist for a watch, feet for footwear.
![STUDIO model picker showing a grid of model faces with gender, ethnicity and body-part filters]()
Match the face to your **product category and target market**. Fine jewellery and a hand model; streetwear and a younger, casual model; a luxury market and a model your customers recognise as one of their own. The closer the model fits the buyer, the more the image sells.
## Generate a face vs. swap a face
Two templates deal with the face itself. They sound similar but do opposite jobs — pick by what you already have.
**Creates a brand-new, consistent face** that doesn't exist yet. Use it when you want a signature model for your brand — one you can bring back for every drop.
**Puts a chosen face onto an existing frame.** Use it when the pose, outfit and scene are already right and you only need to change who's wearing it.
Reach for **Model Face Generation**. It builds a fresh face from scratch — no real person, no likeness rights to clear — and the result stays consistent so you can reuse it. This is how you get a house model that's uniquely yours.
Reach for **Model Face Swap**. The composition is done; you just replace the face with one from the library. Everything else in the frame stays put.
Generation **invents** a face. Swap **replaces** a face on an image you already have. If you don't yet have a frame to swap onto, start with Generation.
## Reuse one face across a whole drop
The point of a consistent face is repetition. Lock in one model and run it through every product in a collection — the same person in every shot reads as one campaign, not a stack of unrelated images.
Generate a new signature face, or pick one from the library and note it.
Use that same model on every product in the drop — each new product image, one consistent face.
Everything auto-saves to your Library, so your house model is there to bring back for the next launch.
Brand consistency comes from **holding the model steady** while the product changes — not from re-picking a new face each time. One face, one campaign.
## Good inputs still matter
A model template is only as good as the product photo you feed it. The AI dresses your model in the product you upload — a clean, correctly-angled input gives a clean result.
Plain background, single product, centered and fully in frame
Sharp and high-resolution — no blur, no crop
See [Product Images](/input-uploads) for the full product-photo checklist before you generate.
## Where to go next
Every template and the one job it does — including all the model templates.
How to prep the product photo, reference image and model selection.
# Platform Overview
Source: https://docs.holala.ai/platform-overview
The three surfaces of holala — Studio, Editor, and Templates — and how a raw product photo becomes a finished, brand-consistent image.
holala is built around three surfaces that work together: **Studio**, where you generate; the **Editor**, where you refine; and **Templates**, the one-job functions that do the actual work. Learn these three and you understand the whole product.
No prompting, no photoshoot. You upload a product photo, pick a Template, and Generate. Everything happens inside these three surfaces.
## The three surfaces
Your workspace. Pick a Template, add your inputs, choose resolution and variations, and Generate.
Refine a result — review variations, polish the winner, and get it export-ready.
The catalog of one-job functions — on-model, background, ghost mannequin, and more.
## How a product photo becomes a finished image
Open [Studio](/studio) and pick the [Template](/templates) for the job you want done.
Upload a product image (required). Add an optional reference image for style, or select a model for on-model templates. See [Product Images](/input-uploads).
Choose 1K for a test or 2K / 4K for final, set how many variations, and Generate. Results save to your Library.
Open a result in the [Editor](/editor) to polish it, then export as JPG, PNG, or WebP.
## Everything saves to your Library
Every generation is stored automatically in your **Library**, so you can compare variations, re-download, or push a finished image back to your store. Connect **My Store** (Shopify, IKAS) to pull products in and send results back.
Browse every one-job function and pick the right one.
What each generation costs, and why 1K is for tests only.
# Tips & Tricks
Source: https://docs.holala.ai/pro-tips
Power-user workflows for getting better holala results while spending fewer credits.
Once you know the basics, these habits are what separate a quick test from a repeatable, scalable workflow. None of this is required — but each one saves credits, time, or both.
The single highest-leverage habit on this page is testing at 1K before committing to a final resolution. Read that one first.
Higher resolution and more variations cost more credits. Instead of generating your first attempt at 4K, run it at 1K first to confirm the Template, inputs, and composition are right.
Once the 1K test looks correct, re-run the same setup at 2K or 4K for the version you'll actually publish. You only pay full price once, on the version you keep. See [Credits](/credits) for exact cost behavior.
A reference image guides style, scene, and lighting — and a reference that works well for one product usually works well for similar products in your catalog.
Save the reference images that produce results you like, and reuse them across a product line. This gets you consistent lighting and mood across a whole collection without re-deciding style every time.
If you're generating on-model shots across a multi-item collection, pick one model from [Model Faces](/model-faces) and reuse the same selection for every item.
This keeps the model consistent across your product grid or lookbook, so the collection reads as one coherent shoot instead of a mix of different faces.
When you have many products to process, don't do them one at a time. Connect My Store (Shopify or IKAS) to pull products directly into holala, then use bulk generation to process them as a batch.
Finished images push back to your store automatically, which removes the manual download-and-upload step entirely.
Every Template is a one-job function — built for a specific kind of output. If a result feels like it's fighting you, the fix is often a different Template, not different settings.
Browse [Templates](/templates) before you start and match the Template to the job, rather than adjusting inputs to force a Template into a job it wasn't built for.
If you're not sure what a Template or model can produce, look through [Inspiration](/inspiration) first. It's a curated library of presets, model faces, and reference scenes you can pull directly into your own generation instead of starting from a blank setup.
## A typical scale-up workflow
Match the Template to the job — don't force one. Browse [Templates](/templates) if you're unsure which fits.
Reuse a saved reference image and a locked [model face](/model-faces) so the whole collection stays visually consistent.
Confirm the composition and styling look right before spending on higher resolution.
Re-run the confirmed setup at final resolution. Check [Credits](/credits) to plan spend across the batch.
For a full catalog, connect My Store and run bulk generation instead of repeating the steps above product by product.
Understand how resolution and variation count affect credit spend.
Browse every Template and what job it's built for.
Pick and lock a model for consistent on-model collections.
A curated library of presets, model faces, and reference scenes.
# Before Generation
Source: https://docs.holala.ai/quickstart
Output quality in Holala is directly tied to the image you upload. Follow the guidelines on this page and you'll significantly reduce the chance of unexpected results.
**Important:** A poor input image causes the AI model to misread your product. Credits are consumed but you won't get the output you expect.
## How to prepare your product image
The product image you upload gives the AI model all the information it has about your product. Any issues in this image will directly affect the output.
Plain or white background\
Single-color, clean surface\
Product centered in frame\
Shot from the angle you want in the output\
Sufficient resolution — not blurry\
Full product in frame, not cropped
Tripods, light stands, or cameras in background\
Complex or patterned background\
Wanting a front view but uploading a side angle\
Part of the product cut off by the frame\
Low resolution or heavily compressed image\
Multiple products in the same frame
**Angle rule:** If you want a front-facing output, your uploaded image must also be front-facing. The AI model cannot generate a front view from a side-angle photo.
### Cropping and background cleanup
Crop your image tightly before uploading. Don't leave excessive empty space around the product. If possible, remove the background beforehand — a plain white or transparent background gives the AI model the cleanest possible input.
## Choosing a reference image
The reference image is your second input and guides the style of the output. The cleaner and more focused your reference, the more predictable the result.
A single, clear style or atmosphere\
Shows the lighting and color tone you want\
A visual from a similar product category
A busy scene with too many elements\
Unrelated category (mismatched product type)\
Low resolution or blurry image\
Part of the product cut off by the frame\
Multiple products in the same frame
## Next step
A generation in Holala is completed in five steps. Follow them in order.
The full input-prep guide, with per-category tips.
# Showcases
Source: https://docs.holala.ai/showcases
Before/after transformations from real brands using holala across apparel, footwear, jewellery, cosmetics, and bags.
A raw product photo in, a brand-consistent marketing image out. Here's what that looks like across categories — no photoshoot, no studio, no prompting required.
Every transformation above started as a single product photo. The product never leaves the frame — holala refines lighting, scene, and presentation around it, without a reshoot.
Browse the curated library of presets, model faces, and reference scenes behind looks like these.
See every Template available and pick the one built for your category.
# Studio
Source: https://docs.holala.ai/studio
Studio is where you generate — pick a template, add your inputs, choose resolution and variations, and click Generate. Every result saves to your Library.
**Studio** is the heart of holala — the surface where every image is made. You don't write prompts here; you pick a **Template**, give it your inputs, and click **Generate**.
Studio shows you exactly which inputs the template you picked needs. Upload what it asks for and Generate — that's the whole loop.
## The Studio loop
Choose the [Template](/templates) that does the job you want — on-model, background, ghost mannequin, and so on. The template is the instruction.
Add your **product image** (always required). Some templates also take an optional **reference image** for style and scene, or a **model selection**. See [Product Images](/input-uploads) and [Model Faces](/model-faces).
Pick **1K** for a quick test or **2K / 4K** for final output, and choose how many variations to generate. More of each costs more [credits](/credits).
Click Generate. Your results appear in seconds and save automatically to your Library.
## What's on the Studio panel
Every one-job function, grouped by what it does. Pick one to open its inputs.
Drop your product photo, add an optional reference, and select a model where the template supports it.
Choose 1K / 2K / 4K and how many versions to generate in one run.
Every generation lands here automatically — compare, re-download, or send back to your store.
Run a cheap **1K** test first to check the composition, then re-generate the winner at **2K or 4K**. It's the simplest way to save credits. [How credits work](/credits).
## From Studio to finished image
Once you like a result, open it in the [Editor](/editor) to refine it, then export as JPG, PNG, or WebP — or push it straight back to your store if you've connected **My Store**.
See every job Studio can do.
Clean inputs are the biggest lever on quality.
# Templates
Source: https://docs.holala.ai/templates
A Template is a one-job function inside Studio. Pick the one that does what you need, upload your inputs, and Generate. No prompt-writing.
A **Template** is a function that does one job. Product-to-On-Model puts your product on a model. White Background drops a clean packshot background. Ghost Mannequin removes the mannequin. You never describe what you want in words — you pick the Template that already does it.
There is **NO PROMPTING**. Choosing the right Template *is* the instruction. A few Templates accept a short optional prompt to nudge the result, but you never have to write one to get an image.
## How a Template works
Each Template card names the exact job it does. Pick one to open its panel.
Every Template asks for a **product image** (required). Some also take an optional **reference image** for style and scene, a **model selection**, or a short prompt. Studio shows you which inputs that Template needs.
Pick 1K / 2K / 4K and how many variations you want, then hit Generate. Results auto-save to your Library.
New to the flow? Start with the [Getting Started](/index) guide, and read [Product Images](/input-uploads) so your first generation comes out clean.
## Templates by category
Put your product on a person — or take the person back out.
Turn a flat or packshot product into a photo worn by a real-looking model.
Remove the mannequin while keeping the fabric drape and silhouette intact.
Keep the product and pose, swap in a different model.
Take the model out of an on-model shot and leave the product alone.
Set the scene and background around your product.
Drop your product into a generated scene or setting.
Build a new background around the product you upload.
Produce a clean, catalog-ready white-background packshot.
Model-free product shots — flat, hanging, or propped — with no cast.
Add views and sharpen the details you don't have a photo for.
Generate additional angles of a product from your hero frame.
Create the side view of a product — packshot, on-model, or lifestyle.
Macro detail shots — fabric weave, stitching, leather grain — from one hero.
Sharpen and upscale an existing image to final-output quality.
Control the model's face.
Generate a fresh, consistent model face for your shots.
Swap the face on an on-model image for a different one.
Browse ready-made faces in the [Model Faces](/model-faces) library.
Templates tuned for product types that need special handling.
Built for the reflections and fine detail of jewellery.
Tuned for children's apparel and products.
Handles fit and skin for lingerie and swimwear.
Move fast across a whole catalog.
Run the same job across many products at once.
Reverse-engineer a clean flat-lay packshot from any on-model frame.
## Every Template lists its own inputs
You don't have to memorize which Template needs what. When you open a Template in Studio, it shows exactly which inputs it requires — product image, optional reference, model selection, or a short prompt. Upload what it asks for and Generate. See [Product Images](/input-uploads) for how to prep each input, and [Credits](/credits) for what each generation costs.
# Templates
Source: https://docs.holala.ai/templates-overview
A Template is a one-job function inside Studio. Choosing the right template IS the instruction — there is no prompting in holala.
Everything holala does, it does through a **Template** — a function that performs one specific job. Put a product on a model, remove a mannequin, drop a clean white background: each is its own template. **Choosing the right one is how you tell holala what you want.** There's no prompt to write.
This is the idea behind the whole product: **no prompting**. You don't describe the result in words — you pick the template that already produces it.
## Why templates instead of prompts
Each template is tuned for one outcome, so you get consistent, professional results without learning how to phrase a prompt.
Open a template and Studio shows exactly what to upload — product image, optional reference, or a model. No guesswork.
## The categories at a glance
Product-to-On-Model, Model Swap, Model Remover, Ghost Mannequin.
Product Placement, Background Generation, White Background, Non-Model Generation.
View / Angle Generator, Side Generation, Closeup Textile, Quality Enhancer.
Model Face Generation and Swap, Jewellery, Kids, Lingerie & Swimwear, Bulk Generation, On-Model to Flat-Lay.
Not sure which to use? Every template names the exact job it does. Browse the full catalog and pick by outcome.
The full catalog, grouped by job, with every template's inputs.
The input rules that make every template work better.
# Background Generation
Source: https://docs.holala.ai/templates/background-generation
Build a new background or scene around your product without a photoshoot, studio, or prompt.
Background Generation replaces your product's backdrop with a new scene, so one photo becomes a full set of on-brand visuals.
## What it does
Upload a product photo and Background Generation builds a new environment around it, from a plain seamless backdrop to a styled lifestyle scene. The product itself stays untouched — only what surrounds it changes. There's nothing to type; the look comes from your inputs, not a text prompt.
## Inputs
* **Product image** (required) — a clean, well-lit shot of the product. See [Product Images](/input-uploads) for prep guidance.
* **Reference image** (optional) — a scene, backdrop, or lighting reference that guides the new background's style.
* **Model selection** (optional) — if the scene should include a person, filter by gender, ethnicity, or body part.
The item to feature. Sharp focus, good lighting, minimal clutter.
Sets the mood, setting, or lighting for the new background.
## How to use it
Select this Template from your Studio library.
Add the required product photo, plus an optional reference image if you have a specific scene in mind.
Choose 1K for a quick test or 2K/4K for final output, then pick how many variations to generate.
Results save automatically to your Library, ready to download as JPG, PNG, or WebP.
## Tips
If your product photo is lit from one side, pick a reference scene with similar lighting so the product doesn't look pasted in.
Run a 1K test to confirm the background direction before spending credits on 2K or 4K variations.
A plain, distraction-free product photo gives the model the cleanest edges to work with when building the new scene around it.
## Related
Need a neutral catalog shot instead? Go clean, not styled.
How to prep product and reference images for the best results.
# Bulk Generation
Source: https://docs.holala.ai/templates/bulk-generation
Run one template across many products at once instead of generating images one by one.
Bulk Generation runs the same template across many products at once, instead of one by one.
## What it does
When you have a full catalog to shoot, generating one product at a time doesn't scale. Bulk Generation applies a single template — the same job, the same settings — to a whole batch of product images in one pass, so a catalog update takes minutes instead of days.
## Inputs
**Required**
* Product images — multiple product photos, uploaded together or pulled from a connected store. See [Product Images](/input-uploads) for shot guidelines.
**Optional**
* Reference image and model selection — applied uniformly across the batch if the chosen template supports them. See [Model Faces](/model-faces).
Connect **My Store** (Shopify or IKAS) to pull products directly into Bulk Generation and push finished images back once they're done — no manual export and re-upload.
## How to use it
Choose multiple product images to upload, or select products from your connected store feed.
Pick the single template to apply across the whole batch — every product in the run uses the same job.
Choose 1K, 2K, or 4K and the number of variations per product. This applies to every item in the batch.
Run the batch. Results save to your Library per product, and push back to My Store automatically if connected.
## Tips
Bulk Generation applies one template uniformly — products shot at different angles or lighting will carry those differences into the output. Standardize your inputs first.
Run 5-10 products first at 1K to confirm the template and resolution choice, then commit the full catalog at 2K or 4K.
If you're refreshing images regularly, a connected store feed saves you from re-uploading products every time.
## Related
Understand how resolution and variation count affect credit use at scale.
Browse every template you can apply across a batch.
# Closeup Textile Generation
Source: https://docs.holala.ai/templates/closeup-textile-generation
Generate macro detail shots of fabric weave, stitching, or leather grain from a single hero product image.
Closeup Textile Generation produces macro detail shots — fabric weave, stitching, leather grain — from the one hero image you already have.
## What it does
This template zooms into the material of your product and renders it as a sharp, close-up detail shot. Instead of shooting a separate macro photo, you upload the hero image you already have and holala extracts and enhances the texture — weave, thread, grain, or stitch line — into its own standalone visual. It's built for the detail shots that sit alongside your main product photo in a listing or lookbook.
## Inputs
**Product image** (required) A sharp, well-lit shot of the product where the material is clearly visible. See [Product Images](/input-uploads) for prep guidelines.
**Reference image** (optional) A close-up style or lighting reference if you want the detail shot to match a specific mood or framing.
This template doesn't use model selection — it works on the material itself, not a person wearing it.
The AI enhances and magnifies the texture already present in your photo — it can't invent a weave or grain the input doesn't show. A blurry or low-resolution source produces a blurry close-up.
## How to use it
From [Studio](/studio), select the Closeup Textile Generation template.
Add the product photo that shows the material you want in macro detail. Optionally add a reference image for framing or lighting style.
Choose 1K for a quick test, or 2K/4K for final output. Pick how many variations to generate — more variations and higher resolution use more credits. See [Credits](/credits).
Results save automatically to your Library, ready to download as JPG, PNG, or WebP.
## Tips
Macro detail has nowhere to hide a soft source. A crisp, high-resolution hero image gives the AI real texture to work from instead of guessing at it.
A tight, well-focused view of the weave, stitching, or grain gives cleaner results than a wide shot where the texture is a small part of the image.
Harsh shadows or glare on the original photo carry into the close-up and can hide the very detail you're trying to showcase.
## Related
Already have a close-up but it needs sharpening? Upscale it to final quality.
Get your product photos ready before generating.
# Ghost Mannequin
Source: https://docs.holala.ai/templates/ghost-mannequin
Remove the mannequin from a garment photo while keeping its hollow, 3D shape and drape.
Ghost Mannequin removes the mannequin from your garment photo while keeping its hollow, three-dimensional shape and drape.
## What it does
This template takes a photo of a garment shot on a mannequin and produces a clean ghost-mannequin image — the mannequin disappears, but the garment keeps its dimensional form, collar structure, and natural drape, just as if it were photographed with the traditional invisible-mannequin technique.
## Inputs
* **Product image** (required) — a photo of the garment on a mannequin, with the shape and drape clearly visible. See [Product Images](/input-uploads) for prep guidelines.
Ghost Mannequin doesn't use a reference image or model selection — it works from your garment shot alone.
The AI refines what your photo shows — it can't invent structure the input doesn't reveal. If the mannequin photo is shot from the front, the result will be a front view.
## How to use it
From [Studio](/studio), select the Ghost Mannequin template.
Add the mannequin shot you want converted.
Choose 1K for a quick test, or 2K/4K for final output. Pick how many variations to generate — more variations and higher resolution use more credits. See [Credits](/credits).
Results save automatically to your Library, ready to download as JPG, PNG, or WebP.
## Tips
These areas define the garment's structure in the final result. If they're obscured or cropped in the source photo, the AI has nothing to work from.
A simple background helps the AI cleanly separate the garment from the mannequin, producing a crisper hollow-shape result.
A closed garment holds its shape better on the mannequin, which carries through to a more natural drape in the output.
## Related
Get your product photos ready before generating.
Have an on-model shot instead of a mannequin? Remove the model and keep the product clean.
# Jewellery Image Generation
Source: https://docs.holala.ai/templates/jewellery-image-generation
A template tuned for the reflections and fine detail of jewellery, from rings to necklaces.
Jewellery Image Generation is built specifically for the reflections, sparkle, and fine detail that jewellery photography demands.
## What it does
Jewellery is small, reflective, and unforgiving of soft detail — general-purpose templates can flatten metal highlights or blur fine settings. This template is tuned for that: it handles gemstone facets, metal reflections, and delicate settings so the result keeps the sparkle and precision that make jewellery photography sell. Upload your piece and, optionally, place it on a model's hand, wrist, or ear.
## Inputs
**Product image** (required) A sharp, close, well-lit shot of the piece. See [Product Images](/input-uploads) for prep guidelines.
**Model selection** (optional) For on-body shots, filter by body part — hand, wrist, or ear — to match the piece. Browse options at [Model Faces](/model-faces).
An optional reference image can also guide the lighting or scene style.
The AI enhances the reflections and detail already present in your photo — it can't invent facets or settings the input doesn't show. A soft or low-resolution source limits how sharp the sparkle can come out.
## How to use it
From [Studio](/studio), select the Jewellery Image Generation template.
Add a sharp product shot of the jewellery. If you want an on-body shot, choose a body part in model selection.
Choose 2K or 4K for final output — jewellery detail needs the resolution. Pick how many variations to generate. See [Credits](/credits).
Results save automatically to your Library, ready to download as JPG, PNG, or WebP.
## Tips
Jewellery detail lives in the facets, settings, and metal grain — a high-resolution, in-focus source photo is the single biggest factor in a good result.
A 1K test won't show the reflections and fine detail clearly enough to judge the result. Go straight to 2K or 4K for jewellery.
Avoid harsh glare or busy reflections in your source shot — a clean, evenly lit piece gives the AI true highlights to work with instead of confusing hotspots.
## Related
Filter by body part to place jewellery on a hand, wrist, or ear.
Need an even tighter macro shot of the setting or texture? Use this template.
# Kids Image Generation
Source: https://docs.holala.ai/templates/kids-image-generation
Generate brand-consistent marketing images for children's apparel and products without a photoshoot.
Kids Image Generation turns a single product photo into polished marketing images built specifically for children's apparel and kids' products.
## What it does
This template is tuned for the proportions, poses, and presentation that children's product photography needs — smaller frames, softer lighting, and age-appropriate styling. Upload a product shot and holala refines it into a clean, brand-ready image without a photoshoot, a studio, or a prompt.
## Inputs
**Required**
* Product image — your kids' apparel or product photo, shot clearly and evenly lit. See [Product Images](/input-uploads) for guidance.
**Optional**
* Reference image — a pose, background, or styling reference for the generation to follow.
* Model selection — choose a child model's gender and ethnicity. See [Model Faces](/model-faces) for details.
holala refines what's already in your photo. It can't invent an angle, detail, or garment feature the input doesn't show.
## How to use it
Open STUDIO and select Kids Image Generation.
Add your product image, plus an optional reference image or model selection if this product needs one.
Choose 1K to test the composition quickly, or 2K/4K for final output. Pick how many variations you want.
Run the template. Results save automatically to your Library, ready to download as JPG, PNG, or WebP.
## Tips
Children's apparel tends to have smaller details — snaps, prints, trims. Even lighting on the input photo keeps those details crisp in the output.
If you're generating a full collection, keep the model selection consistent across products so the set reads as one campaign.
Run a 1K pass first to confirm framing and styling, then regenerate at 2K or 4K once you're happy with the result.
## Related
Learn how model selection works across templates.
Prep your product photos for the best results.
# Lingerie & Swimwear Generation
Source: https://docs.holala.ai/templates/lingerie-swimwear-generation
Generate on-model lingerie and swimwear images with accurate fit, drape, and skin rendering.
Lingerie & Swimwear Generation handles the fit, drape, and skin rendering that lingerie and swimwear photography demands.
## What it does
Lingerie and swimwear are unforgiving categories — fabric tension, strap placement, and skin tone all have to look right or the image feels off. This template is tuned for that: it reads your product photo and renders it on-model with realistic drape and natural skin, no photoshoot required.
## Inputs
**Required**
* Product image — your lingerie or swimwear piece, laid flat or shot on a form. See [Product Images](/input-uploads) for how to shoot it.
**Optional**
* Reference image — a pose or angle reference for the final shot.
* Model selection — choose gender, ethnicity, and body part. See [Model Faces](/model-faces).
The AI refines what's in your photo — it can't invent a strap, seam, or cut that isn't visible in the input.
## How to use it
Open STUDIO and select Lingerie & Swimwear Generation.
Add a clear, evenly lit shot of the piece, plus a reference image or model selection if you need a specific look.
Use 1K for a quick fit check, or 2K/4K for final marketing output. Set the number of variations you want.
Run the template. Finished images land in your Library, downloadable as JPG, PNG, or WebP.
## Tips
A smooth, well-lit product photo gives the model the clearest read on drape and fabric tension — wrinkles or shadows in the input can carry through to the output.
Set body part and model attributes to keep looks consistent across a collection, especially if you're generating multiple pieces in the same drop.
Fit and drape can shift slightly between variations. Running 2-3 per product gives you options to pick the most flattering result.
## Related
See how gender, ethnicity, and body part selection work.
Avoid the input errors that most often affect fit and drape.
# Model Face Generation
Source: https://docs.holala.ai/templates/model-face-generation
Generate a fresh, consistent model face you can reuse across a whole collection or campaign.
Model Face Generation creates a brand-new, consistent model face you can reuse across every product in a collection.
## What it does
This template builds a fresh face from scratch — no real person, no likeness rights to clear — and keeps it consistent so the same face can appear across every shot in a drop. Use it when you want a signature model that's uniquely yours, one you can bring back season after season instead of re-casting each time.
Generate **invents** a new face. If you already have an on-model image and just want to change who's wearing it, use [Model Face Swap](/templates/model-face-swap) instead.
## Inputs
**Product image** (required) A clear, well-lit shot of the product the generated model will be shown with. See [Product Images](/input-uploads) for prep guidelines.
**Model selection** (required) Filter by gender, ethnicity, or body part to set the direction for the new face. Browse the filters at [Model Faces](/model-faces).
An optional reference image can also guide pose, scene, or lighting style.
## How to use it
From [Studio](/studio), select the Model Face Generation template.
Add your product photo, then choose gender, ethnicity, or body part to steer the new face. Optionally add a reference image for pose or scene.
Choose 1K for a quick test, or 2K/4K for final output. Pick how many variations to generate — more variations and higher resolution use more credits. See [Credits](/credits).
Results save automatically to your Library, ready to download as JPG, PNG, or WebP.
## Tips
Once you generate a face you like, keep using that same result across every product in the collection instead of generating a new one each time — that repetition is what reads as one campaign.
Run a small batch at 1K to compare a few face options, then pick your favorite and regenerate your final product shots at 2K or 4K.
The closer the generated face fits your product category and target market, the more natural the campaign feels. See [Model Faces](/model-faces) for guidance on picking filters.
## Related
Understand model selection and how to keep one face consistent across a drop.
Already have a frame you like? Swap in a different face instead of generating a new scene.
# Model Face Swap
Source: https://docs.holala.ai/templates/model-face-swap
Swap the face on an existing on-model image for a different one, without reshooting the pose or scene.
Model Face Swap replaces the face on an existing on-model image with a different one, while everything else in the frame stays put.
## What it does
This template takes an on-model photo you already have — the pose, outfit, and scene are already right — and swaps in a different face from the model library. Nothing else in the image changes. Use it when the composition is done and you only need to change who's wearing it, instead of generating an entirely new shot.
Swap **replaces** a face on an image you already have. If you don't yet have an on-model frame to start from, use [Model Face Generation](/templates/model-face-generation) instead.
## Inputs
**Product image** (required) The existing on-model photo whose face you want to replace. See [Product Images](/input-uploads) for prep guidelines.
**Model selection** (required) Choose the replacement face by filtering gender, ethnicity, or body part. Browse the library at [Model Faces](/model-faces).
The AI replaces the face while preserving the rest of the frame — it can't fix a pose or crop issue elsewhere in the source photo. Start from a clean on-model shot for the best result.
## How to use it
From [Studio](/studio), select the Model Face Swap template.
Add the existing photo, then filter the model library by gender, ethnicity, or body part to pick the replacement face.
Choose 1K for a quick test, or 2K/4K for final output. Pick how many variations to generate — more variations and higher resolution use more credits. See [Credits](/credits).
Results save automatically to your Library, ready to download as JPG, PNG, or WebP.
## Tips
The swap works best when the original face is clearly visible, well-lit, and not turned at an extreme angle — that gives the AI a clean area to replace.
If you're standardizing a whole drop on one face, swap that same model into every existing on-model shot for a consistent campaign look.
If you're swapping the same face across many images, confirm the match looks right on one image at 1K before running the rest at 2K or 4K.
## Related
Browse the model library and its gender, ethnicity, and body-part filters.
Don't have an on-model frame yet? Generate a fresh face and shot from scratch.
# Model Remover
Source: https://docs.holala.ai/templates/model-remover
Take the model out of an on-model shot and leave a clean product image behind.
Model Remover takes the model out of an on-model shot and leaves the product clean.
## What it does
This template starts from a photo of a model wearing or holding a product, and removes the model entirely — producing a clean, standalone product image. It's built for turning existing on-model photography into flat lay or packshot-style images without a separate shoot.
## Inputs
* **Product image** (required) — a photo of the model wearing or holding the product you want to isolate. See [Product Images](/input-uploads) for prep guidelines.
Model Remover doesn't use a reference image or model selection — it works from your on-model shot alone.
The AI refines what your photo shows — it can't invent parts of the product that the model's pose hides. If a sleeve or seam is covered by the model's arm, that detail won't appear in the result.
## How to use it
From [Studio](/studio), select the Model Remover template.
Add the shot you want the model removed from.
Choose 1K for a quick test, or 2K/4K for final output. Pick how many variations to generate — more variations and higher resolution use more credits. See [Credits](/credits).
Results save automatically to your Library, ready to download as JPG, PNG, or WebP.
## Tips
Any part of the product hidden behind the model's body or hair won't be reconstructed. Choose the shot with the least overlap.
A straightforward pose — standing, product held out or worn plainly — gives a cleaner isolation than a dynamic or cropped pose.
Zoom into where the product met the model's skin or hair in the original shot, since that's where results benefit most from a 4K pass.
## Related
Have a mannequin shot instead of an on-model shot? Remove the mannequin and keep the garment's shape.
Drop your isolated product into a new generated scene or setting.
# Model Swap
Source: https://docs.holala.ai/templates/model-swap
Keep the product and pose in an on-model photo, and replace the model with a different one.
Model Swap keeps your product and pose exactly as shot, and replaces the model with a different one.
## What it does
This template takes an existing on-model photo and swaps in a new model, while keeping the product and pose unchanged. It's the fastest way to reuse a shoot you already have — showing the same product on a different gender, ethnicity, or look — without booking a new session.
## Inputs
**Product image** (required) Your existing on-model photo, with the product and pose clearly visible. See [Product Images](/input-uploads) for prep guidelines.
**Model selection** (required for the swap) Filter by gender, ethnicity, or body part to choose the new model. Browse options at [Model Faces](/model-faces).
An optional reference image can also guide styling details you want carried into the result.
The AI refines what's visible in your source photo — it keeps the pose and product as shot, so an input at a three-quarter angle will produce a three-quarter result.
## How to use it
From [Studio](/studio), select the Model Swap template.
Add the shot whose product and pose you want to keep, then pick the new model from the model selection filters.
Choose 1K for a quick test, or 2K/4K for final output. Pick how many variations to generate — more variations and higher resolution use more credits. See [Credits](/credits).
Results save automatically to your Library, ready to download as JPG, PNG, or WebP.
## Tips
A pose with the product clearly visible and unobscured gives the AI the clearest read on fit and shape to carry onto the new model.
If your source photo is a close-up crop — hands, torso, feet — narrow the model selection to that body part for a more consistent swap.
Small differences in lighting and angle interpretation are normal. Generating 2-3 variations gives you options to pick the most natural result.
## Related
Browse and filter the models available for on-model templates.
Don't have an on-model shot yet? Start from a flat lay or packshot instead.
# Non-Model Generation
Source: https://docs.holala.ai/templates/non-model-generation
Create model-free product shots — flat, hanging, or propped — with no person in frame.
Non-Model Generation produces model-free product shots — flat lay, hanging, or propped — with no person in the frame.
## What it does
Upload a product photo and Non-Model Generation presents it on its own, styled as a flat lay, hung on a rail or hook, or propped against a surface. It's built for products that need to be shown clearly without a person wearing or holding them, keeping full attention on the item itself.
## Inputs
* **Product image** (required) — the item to feature, shown as flat, hanging, or propped. See [Product Images](/input-uploads) for prep guidance.
* **Reference image** (optional) — a styling reference for how the product should be presented (surface, angle, props).
A clear shot of the product, ideally already flat or hung if possible.
Guides the styling — surface, props, or presentation angle.
## How to use it
Select this Template from your Studio library.
Add the required product photo, plus an optional reference image if you want a specific styling.
Choose 1K to test the presentation style, then 2K or 4K for final output.
Results save automatically to your Library, ready to download as JPG, PNG, or WebP.
## Tips
Use a reference image to signal flat lay versus hanging versus propped — the style comes from what you show it, not a description.
For flat lay results, an input photo where the product is already laid flat (not bunched or folded) gives the cleanest output.
A busy reference scene can pull focus from the product — simple references tend to produce cleaner, more commercial results.
## Related
Need a plain catalog packshot instead of a styled flat lay?
How to prep product and reference images for the best results.
# On-Model to Flat-Lay
Source: https://docs.holala.ai/templates/on-model-to-flat-lay
Reverse-engineer a clean flat-lay packshot from an on-model product photo.
On-Model to Flat-Lay reverse-engineers a clean flat-lay packshot from an on-model frame.
## What it does
Sometimes the only photo you have of a product is on a model, but you need a clean packshot for a listing or catalog page. This template reads the on-model image and produces a flat, product-only version — no photoshoot or restyling needed.
## Inputs
**Required**
* Product image — an on-model photo showing the product clearly. See [Product Images](/input-uploads) for what makes a usable source frame.
**Optional**
* Reference image — a flat-lay style or background reference for the output.
The output can only show what the on-model photo shows. If part of the product is hidden by a pose or angle, that detail won't appear in the flat-lay.
## How to use it
Open STUDIO and select On-Model to Flat-Lay.
Add the source image, plus a reference image if you want a specific flat-lay style.
Use 1K to check the result quickly, or 2K/4K for a catalog-ready packshot. Set how many variations to generate.
Run the template. The flat-lay result saves to your Library, ready to download as JPG, PNG, or WebP.
## Tips
Pick a source photo where the product is fully visible and not obscured by the pose, hair, or accessories — that's what the flat-lay will be built from.
A straight-on frame gives the cleanest flat-lay proportions. Angled or side-facing shots can distort the reverse-engineered layout.
Confirm the flat-lay captures the product accurately before spending credits on 2K or 4K output.
## Related
Learn what makes a strong source photo for this template.
See the input issues that most often affect flat-lay accuracy.
# Product Placement
Source: https://docs.holala.ai/templates/product-placement
Drop your product into a generated scene or setting for editorial-style marketing images.
Product Placement drops your product into a generated scene or setting, without a location shoot.
## What it does
This template takes your product photo and places it into a new environment — a lifestyle setting, a styled backdrop, or an editorial-style scene — built to match your brand. It's for sellers who need context and atmosphere around a product without booking a location or set.
## Inputs
**Product image** (required) A clear shot of the product you want placed into a scene. See [Product Images](/input-uploads) for prep guidelines.
**Reference image** (optional) Guides the style, scene, or lighting of the generated setting. Use this to match a mood board or a previous campaign look.
Model selection isn't used in this template — it's focused on the product and its setting, not a person.
The AI refines what your product photo shows — it can't invent an angle or surface the input doesn't reveal. A packshot with no visible base won't gain one in the placed scene.
## How to use it
From [Studio](/studio), select the Product Placement template.
Add your product photo, and optionally a reference image to guide the scene's style or lighting.
Choose 1K for a quick test, or 2K/4K for final output. Pick how many variations to generate — more variations and higher resolution use more credits. See [Credits](/credits).
Results save automatically to your Library, ready to download as JPG, PNG, or WebP.
## Tips
If you're building a set of images for one campaign, reuse the same reference image across generations to keep lighting and mood consistent.
A product photographed straight-on sits more naturally in a scene composed for a straight-on view. Consider the setting before you shoot the source image.
Since the setting is generated, running several variations gives you a range of environments to choose the strongest one from.
## Related
Explore editorial-style looks built for campaign and lookbook imagery.
Start from a clean, model-free product image before placing it into a scene.
# Product-to-On-Model
Source: https://docs.holala.ai/templates/product-to-on-model
Turn a flat or packshot product photo into a realistic on-model image without a photoshoot.
Product-to-On-Model puts your flat-lay or packshot product onto a realistic-looking model — no photoshoot required.
## What it does
This template takes a product image with no model in it — a flat lay, a ghost mannequin shot, or a simple packshot — and places that exact product onto a model. It's built for sellers who have clean product photography but no on-model imagery to show fit, drape, or scale.
## Inputs
**Product image** (required) A clear, well-lit shot of the product with no model in it. Flat lays and packshots work best. See [Product Images](/input-uploads) for prep guidelines.
**Model selection** (optional) Filter by gender, ethnicity, or body part to match the model to your brand or campaign. Browse options at [Model Faces](/model-faces).
An optional reference image can also guide the pose, scene, or lighting style if you want the result to match a specific look.
The AI refines what's visible in your product photo — it can't invent an angle your input doesn't show. A photo shot straight-on will produce a straight-on result.
## How to use it
From [Studio](/studio), select the Product-to-On-Model template.
Add your flat lay or packshot. Optionally add a reference image for pose or scene, and pick a model from the model selection filters.
Choose 1K for a quick test, or 2K/4K for final output. Pick how many variations to generate — more variations and higher resolution use more credits. See [Credits](/credits).
Results save automatically to your Library, ready to download as JPG, PNG, or WebP.
## Tips
Even lighting with no harsh shadows gives the AI the clearest read on color, texture, and shape — all of which carry over to the on-model result.
If you use a reference image for pose or scene, pick one with a camera angle close to your product photo. Big mismatches make the result look less natural.
Confirm the model, pose, and framing look right at 1K, then regenerate your final variations at 2K or 4K.
## Related
Browse and filter the models available for on-model templates.
Already have an on-model shot? Swap in a different model while keeping the pose.
# Quality Enhancer
Source: https://docs.holala.ai/templates/quality-enhancer
Sharpen and upscale an existing image to final-output quality, ready for print, ads, or your storefront.
Quality Enhancer sharpens and upscales an existing image so it's ready for final output — no re-shoot required.
## What it does
This template takes an image that's already close to what you want — a past generation, an existing product photo, anything a little soft or low-resolution — and enhances it into a sharp, high-resolution final. It doesn't restyle or restage anything; it improves the image you already have. It's the fastest path from "good enough for a test" to "ready to publish."
## Inputs
* **Product image** (required) — the image you want sharpened and upscaled. See [Product Images](/input-uploads) for prep guidelines.
Quality Enhancer doesn't use a reference image or model selection — it works from the single image you upload.
The AI sharpens and upscales what's already in the frame — it can't add detail that was never captured or fix a badly cropped composition. Very low-resolution or heavily compressed sources have less to work with.
## How to use it
From [Studio](/studio), select the Quality Enhancer template.
Add any existing image — a 1K test result, an older product photo, or any source that needs sharpening.
Choose 2K or 4K for a true final-output upscale. Pick how many variations to generate — more variations and higher resolution use more credits. See [Credits](/credits).
Results save automatically to your Library, ready to download as JPG, PNG, or WebP.
## Tips
Run your other templates at 1K to dial in composition, model, or scene, then send the keeper through Quality Enhancer at 2K or 4K instead of re-generating from scratch at high resolution.
Since the job is to reach final-output quality, generating at 1K here defeats the purpose. Go straight to 2K or 4K.
Quality Enhancer sharpens detail — it won't recrop a badly framed shot or invent an angle the source doesn't show. Fix composition issues at the source, then enhance.
## Related
Need a macro detail shot instead of a full-image upscale? Use this template.
See how resolution and variations affect what a generation costs.
# Side Generation
Source: https://docs.holala.ai/templates/side-generation
Create the side view of a product as a packshot, on-model, or lifestyle shot.
Side Generation creates the side view of your product, as a packshot, on-model, or lifestyle shot.
## What it does
Upload a product photo and Side Generation produces a dedicated side-view image, in whichever presentation you need: a plain packshot, on a model, or set in a lifestyle scene. It's built for the specific case where a side profile — not a front or three-quarter view — is the shot you're missing.
## Inputs
* **Product image** (required) — the shot you're generating a side view from. See [Product Images](/input-uploads) for prep guidance.
* **Reference image** (optional) — guides the presentation style: packshot styling, on-model context, or a lifestyle scene.
* **Model selection** (optional) — if the side view should be on-model, filter by gender, ethnicity, or body part.
The source shot the side view will be generated from.
Needed only if the side view should show the product on a person.
## How to use it
Select this Template from your Studio library.
Add the required product photo, plus a reference image and model selection if relevant to your desired result.
Choose 1K to check the side profile, then 2K or 4K for final output.
Results save automatically to your Library, ready to download as JPG, PNG, or WebP.
## Tips
The AI refines what's already visible in your photo — it can't invent a side profile the input never showed. If your source image is strictly front-facing with no side detail visible, the generated side view will be less accurate than if some side contour is already present.
A reference image is the fastest way to signal which presentation you want, since there's no prompt to describe it.
If you want the side view worn by a person, set gender, ethnicity, or body part filters so the result matches your brand's model casting.
## Related
Need more angles beyond just the side profile? Start there.
Browse model options for on-model side-view results.
# View / Angle Generator
Source: https://docs.holala.ai/templates/view-angle-generator
Generate additional angles of a product from a single hero photo.
View / Angle Generator extends one hero photo into additional angles of the same product.
## What it does
Upload a single product photo and View / Angle Generator produces additional views of that product — think three-quarter, side, or back angles that complement your hero shot. It's the fastest way to build a multi-angle gallery from one photoshoot, without re-shooting the product from every side.
## Inputs
* **Product image** (required) — your hero shot. See [Product Images](/input-uploads) for prep guidance.
* **Reference image** (optional) — an angle or framing reference showing the view you want generated.
The hero photo the additional angles will be generated from.
Shows the specific angle or framing you're aiming for.
## How to use it
Select this Template from your Studio library.
Add the required product photo, plus an optional reference image showing the target angle.
Choose 1K to preview the new angle, then 2K or 4K for final output.
Results save automatically to your Library, ready to download as JPG, PNG, or WebP.
## Tips
The AI refines detail already present in your photo — it can't invent an angle or feature the input doesn't show. A front-facing hero shot can't produce an accurate back view; if you need a specific angle, start with an input that already shows at least part of it.
A reference image showing the target framing (three-quarter, side, back) gives more predictable results than relying on the hero shot alone.
Since angle accuracy is the main variable here, a quick 1K pass lets you confirm the new view looks right before spending credits on 2K or 4K variations.
## Related
Need specifically a side view? This Template is purpose-built for that.
What your hero shot needs to show for accurate new angles.
# White Background
Source: https://docs.holala.ai/templates/white-background
Turn any product photo into a clean, catalog-ready white-background packshot in minutes.
White Background produces a clean, consistent packshot on pure white — the standard shot for catalogs and marketplaces.
## What it does
Upload a product photo and White Background strips the existing backdrop, replacing it with a neutral, evenly lit white surface. The product's shape, color, and detail are preserved while the background is normalized, so every item in your catalog looks like it was shot in the same studio, even if it wasn't.
## Inputs
* **Product image** (required) — the item you want isolated on white. See [Product Images](/input-uploads) for prep guidance.
* **Reference image** (optional) — not typically needed for this Template, since the output background is standardized white rather than styled.
A clear, evenly lit shot with the full product visible.
Add a model filter only if the product is shown worn or held.
## How to use it
Select this Template from your Studio library.
Add the required product photo. No reference image needed for a standard white packshot.
Choose 2K or 4K for catalog-ready output, and set the number of variations you need.
Results save automatically to your Library, ready to download as JPG, PNG, or WebP.
## Tips
A product photo with harsh shadows or uneven lighting will carry some of that unevenness into the final packshot. Flat, diffused lighting gives the cleanest result.
Marketplaces and print catalogs often require higher resolution — 4K keeps edges and fine detail sharp when zoomed in.
If you're processing a full catalog, group products shot from the same angle for a more consistent set of results.
## Related
Want a styled scene instead of plain white? Build a full backdrop.
See how resolution and variation count affect credit usage.