Collages on IMG AI

Photo Collage Maker

The IMG AI photo collage maker puts two to six photos into one finished image, each in its own cell. Nothing is regenerated: every photo keeps its original colour and content inside the part of it you keep. Choose a grid that matches your count, set the spacing, nudge a focal point so the important part is not cropped, and download a PNG at 1600 × 1600 or 1600 × 2000.

2–6 photosSquare or 4:51600 px PNGRuns in your browser

Build your collage

Add 2–6 photos. Everything happens in this browser tab: no upload, no account, no credits.

0 / 6 photos

LayoutAdd 2 photos to choose a layout

Shape

Spacing24 px at 1600px

Spacing sits between cells only; the collage keeps its full frame. Zero leaves no line at all.

Preview

1600 × 1600 PNG

Add 2 more photos to build a collage.

Three-cell collage comparing a red, a green, and a blue product variant side by side in equal-width cells on a white background

Three product variants, one comparison grid

Input: Three product shots of the same bottle in red, green, and blue, each photographed at a slightly different distance.
Settings: Three columns · square 1600 × 1600 · spacing 24 px · white background
Output: The three bottles share the same height and the same framing inside their cells, so colour is the difference the eye catches first. Nothing is recoloured, and the white seam between cells is the only pure white in the image.

A catalogue page needs the variants judged against each other, not one at a time. Three equal columns give every cell the same width and the same crop rule, so each bottle is framed the same way and a colour difference is the first thing the eye catches. A thin seam is left in, because spacing makes three separate shots read as three separate products instead of one merged image.

Four-photo collage in a two by two grid showing a coastline, a street, a meal, and a detail shot on a white background

Four travel photos as a 2 × 2 mood board

Input: Four photos from one trip: two landscape, one portrait, one square detail shot.
Settings: 2 × 2 grid · square 1600 × 1600 · spacing 24 px · four mixed orientations
Output: Four equal squares with no letterboxing anywhere. The portrait is cropped, not squashed, and the sky in the first photo still reads as sky.

Mixed orientation is the normal case, not the exception. A 2 × 2 grid gives each photo the same square cell, and the cover crop trims whichever side has room to spare — a landscape loses its edges, a portrait loses its top and bottom, and neither is stretched. The portrait’s focal point is moved up so the face survives the trim.

Collage with a wide product banner across the top showing unbroken lettering, above two detail photographs on a white background

A wide banner with the lettering kept whole

Input: One 2:1 banner carrying a product wordmark, plus a tall close-up of the label and a square texture shot.
Settings: Main cell across the top + two below · square 1600 × 1600 · spacing 24 px
Output: The wordmark is complete inside the main cell, edge to edge. The tall label shot below keeps its full height because the side cell is portrait too.

A wide banner carrying a wordmark is the hardest thing to crop: a few percent off each side and the product name loses a letter. The banner takes the main cell across the top, which spans the full width, so a 2:1 source into a 2:1 cell is barely trimmed at all. Its focal point stays at the middle, which is where the lettering sits, and the two remaining photos take the cells underneath.

What stays unchanged in a photo collage?

Every cell is filled independently, so a photo is never blended with its neighbours, never recoloured, and never re-rendered. The tool scales the source by the larger of the two ratios between the cell and the photo, then crops the overflow. That is the same rule behind object-fit: cover: the cell is always full, and the aspect ratio is never distorted. Whatever falls inside the crop comes out with the colour and content it went in with.

What does change is the framing. A photo whose shape does not match its cell loses the edges that do not fit — a landscape loses its sides in a square cell, a portrait loses its top and bottom. The focal point decides which part survives, which is why the control sits on every photo rather than on the collage: a label near the bottom of a tall bottle should be able to stay in frame while the photo beside it stays centred.

The export is a new file, not a stack of layers. Once the PNG is written, the arrangement is fixed, so keep the original photos and re-export if you need a different grid rather than trying to edit the result.

How do I rearrange photos without regenerating them?

Rearranging a collage is a layout operation, so nothing here asks a model to draw anything. Every photo stays the same file, with the same pixels, in the same browser tab, while the arrangement changes around it.

That matters when a grid looks wrong on the first try. In a generative workflow a new arrangement means a new render, and a new render means new faces, new lighting, and new lettering. Here a swap is a list reorder: the two photos trade cells, each one carrying its own focal point with it, and every other cell keeps exactly the pixels it already had.

Add 2 to 6 photos
Choose JPG, PNG, or WebP files up to 20 MB each. The file is read in the browser and never uploaded, so the picker can accept a batch and still tell you by filename which file was skipped and why.
Pick the grid for that count
Only layouts that hold the current number of photos are offered, so there are no empty cells and no template that asks for a seventh photo. Switching grids re-flows the same photos and never drops one.
Swap cells by drag or by button
Drag a row to reorder it, or use the two selects and the swap button when dragging is awkward — on a phone, for instance. A swap moves a photo and its focal point together, because the focal point belongs to the photo rather than to the cell.
Set the spacing and the focal points
Spacing sits between cells only, so the collage always keeps its full outer frame. Drag the slider to zero for a seamless grid. Click or nudge each photo’s focal pad to choose which part of it the cell keeps.
Download one finished image
Export a 1600 × 1600 square or a 1600 × 2000 four-by-five PNG. The button stays disabled, with the reason printed underneath, until the selection is legal and every photo has finished decoding.

Where a local collage beats a generated one

A collage is the right output whenever the individual photographs have to remain recognisable as themselves. These are the jobs that come up most.

Product variant comparisons

Place the same product in three colours or three sizes side by side at an identical width. A buyer compares the products rather than the photography, which only works if every cell is cropped the same way.

Before-and-after and progress grids

A two-column collage with the same crop on both sides is the clearest way to show a renovation, a restoration, or a fitness change. The spacing tells the eye these are two moments, not one wide photo.

Listing and marketplace galleries

Marketplaces often allow one image per listing, so a 2 × 2 grid of angles packs four views into the slot. Set the spacing to zero for a single seamless sheet, or leave a thin gutter so each angle reads separately.

Mood boards and trip recaps

Four to six photos of a place, a palette, or a project become one image you can drop into a deck or a post without a design tool. Every photo keeps its own colour, which is the point of a mood board.

Proof sheets for a client

A five- or six-cell grid shows a whole set of options in one screen. The client picks from the sheet, and the file you send is the same image you were looking at.

Photo collage maker vs template suites vs desktop publishing vs the AI image combiner

Four routes produce something that looks like a collage, and the difference that matters is what each one does to the pixels you started with. Each link below stays on this site.

IMG AI photo collage maker

A finished collage from photos you already have, when the originals must come out unchanged and you do not want an account to get one image.

Add 2 to 6 photos, choose a grid that matches the count, swap cells by drag or by button, set the spacing and a focal point per photo, then download a 1600 px PNG. Everything runs in the browser tab.

The trade-off: The layouts are geometric grids. There is no text tool, no sticker library, no decorative frame, and no saved project — the PNG is the deliverable.

Big template suites

Decorative collage layouts with headlines, stickers, brand kits, and a team that keeps editing the file afterwards.

Start from a large template library, drop photos into frames, add type and decoration, then export. Templates cover shapes, borders, and seasonal layouts that a geometric grid will never offer.

The trade-off: A subscription for the useful features, a learning curve for the editor, and an account before the first collage. Photographs are usually placed unchanged, but the value is in the decoration rather than in the crop control.

Desktop publishing tools

Print work where a designer needs millimetre placement, bleed, CMYK, and a document that stays editable across sessions.

Set a document, place each image in a frame, set the crop by hand, align the gaps numerically, and export for print or screen.

The trade-off: Full control at the cost of manual work per collage: every frame, gap, and crop is a decision, and the same grid has to be rebuilt for the next set of photos.

The AI image combiner on IMG AI

A new picture that does not exist yet, built from two or three photos — two people in one room, a product in a new setting.

Upload two or three photos, give each one a role in a written brief, and let a model render one coherent scene that relights and re-perspectives the sources.

The trade-off: The output is generated, so faces, labels, and source colours can drift. That is the point when you want a new scene and the wrong tool when the original photographs have to survive intact.

Open the AI image combiner

Pick by what the deliverable has to prove. If a buyer has to recognise each original photo, a collage is the honest format and the grids here are enough. If a client needs decoration and type around the photos, a template suite earns its subscription. If the deliverable is a scene that never existed, use the combiner and expect it to redraw what you gave it.

The last row is the comparison worth reading twice, because it is the one inside this site: the AI image combiner fuses several photos into one generated image. It is the better tool when you want a new scene, and the wrong one when the point of the image is that the photographs are unchanged.

Photo collage maker vs AI image combiner vs AI product photography

Three IMG AI tools that all start from photographs and end somewhere different.

The problem	Photo Collage Maker	AI Image Combiner	AI Product Photography
What you get	One image with every photo unchanged in its own cell	One newly generated image built from two or three photos	One restyled product shot from a single source photo
Source photos	2 to 6	2 or 3	1
Best for	Variant comparison, mood boards, proof sheets, listing galleries	People together, subject in a new scene, product composition	A polished commercial scene around one product
Output	1600 × 1600 or 1600 × 2000 PNG	A 1K generated image in 4:3 or 1:1	A generated product scene
Account and cost	No account, no credits; runs in the browser	Sign-in and credits per generated image	Sign-in and credits per generated image

Your photos stay in your browser

The collage is composed on a canvas inside this tab. No photo is uploaded, no request leaves the page, and no account is involved — closing the page discards the arrangement, so download the PNG before you do. Only use photos you own or have permission to publish, including permission from anyone shown in them.

Specifications

Photos per collage: 2 to 6
Layouts: 3 per photo count, 17 grids in total
Export sizes: 1600 × 1600 square, 1600 × 2000 at 4:5
File formats: JPG, PNG, WebP
Maximum file size: 20 MB per photo
Working pixel budget: 80 megapixels across the selection
Spacing: Between cells only, 0 to 64 px at 1600 px wide
Crop rule: Cover: scale to fill, then crop; never stretched
Output: One PNG, 8-bit RGB
Account, credits, upload: None: the collage is composed locally

Photo Collage Maker FAQ

How many photos can I put in a collage?

Between two and six. Each count has several layouts: two photos can sit side by side, stacked, or as one large cell with a smaller companion; six photos can use a 3 × 2 grid, a 2 × 3 grid, or a feature cell with five smaller cells. Adding a seventh is refused with a message naming the file, and the first six stay exactly where they are.

Why is my photo cropped?

Every cell is filled the way object-fit: cover fills a box, so the photo is scaled until it covers the cell and the overflow is trimmed rather than stretched or letterboxed. A landscape loses its sides in a square cell; a portrait loses its top and bottom. Move the focal point on that photo to choose which part the cell keeps.

Can I change the crop per photo?

Yes. Each photo carries its own focal point, set by clicking or dragging on its focal pad or by nudging it with the arrow buttons. The focal point travels with the photo, so swapping two cells moves each photo and its crop together instead of transferring the old crop to a different picture.

Does a collage reduce image quality?

It can, and it is worth being precise about where. Each cell shows a cropped copy of its photo scaled to the export size, so a small photo dropped into a large cell is upscaled and will look softer than the original. Cell dimensions also differ inside one grid: in a two-by-two layout each cell is half the width of the canvas, so a photo needs roughly 800 pixels on its short side to fill a 1600 px square export at native resolution. Photos at or above that size stay sharp; smaller ones are enlarged.

What size should a collage be?

Start from where it will be published. A square 1600 × 1600 export suits feed posts, listing galleries, and anything that will be shown in a grid. The 1600 × 2000 four-by-five export suits portrait posts and print handouts at small sizes. Both are PNG, which keeps flat colour and hard grid edges clean.

Is this collage maker free?

Yes. The collage is composed in your browser, so there is no generation request, no credit cost, and no account. Files never leave the tab, which also means closing the page discards the arrangement — download the PNG before you close it.

Move to the workflow that matches the image you actually need.

Fuse several photos into one generated scene Turn one product photo into a commercial scene Remove an object from a finished photo Replace a background behind one subject

Make the collage

Add two to six photos, pick a grid, and download one finished image. Nothing is generated, so nothing in your photos changes.

Open the collage maker