Is this a real photo mosaic or just a collage?

It is a colour-matched mosaic. The target is sampled into cells and each cell receives a tile photo selected using measured colour distance plus the chosen diversity and variation controls.

How many tile photos can I use?

You can add two to 100 tile photos. Generation can repeat them across thousands of cells, and a more varied library normally produces better matches.

Will every uploaded photo appear in the mosaic?

Not necessarily. Selection is based on target colours, diversity and variation. A tile whose average colour is unlike every target region may remain unused; the result reports how many unique tiles were selected.

How do I reduce repeated photos?

Add more tile photos and increase Tile Diversity. Some reuse remains necessary whenever the mosaic has more cells than library images.

What do colour tint and target blend do?

Tile tint overlays each cell's target colour on its selected photo. Target blend overlays the overall target after tile placement. Both improve long-distance recognition but can make individual tile photography less natural.

Does the tool recognize faces or subjects?

No. Matching uses average colour, not facial recognition, object detection or semantic AI. Subject-aware placement is not claimed.

Are my photos uploaded to ToolsOnDuty?

No. Image analysis, matching, rendering, project generation and export are designed to happen locally in the browser.

Why can a 4000-pixel mosaic take longer?

Large outputs and dense grids require thousands of tile selections, clips and draws in browser memory. Generation yields periodically and includes a Cancel control, but speed still depends on the device and source images.

Does the project file contain my photographs?

Yes. The target and tile library are embedded so the mosaic can be regenerated later. Treat the project as sensitive and avoid sharing it when the source photos are private.

Is the export ready for large-format professional printing?

It provides the reported RGB pixel dimensions, not guaranteed physical dimensions, DPI metadata, CMYK conversion, bleed or a printer-approved proof. Confirm production requirements separately.

Can I publish a mosaic made from other people's photos?

Only when you have the required copyright, privacy, model-release and commercial permissions for the target and every tile image. A mosaic does not remove those rights.

Image Tools

Photo Mosaic Maker

Build a real colour-matched photo mosaic from one target image and up to 100 personal tile photos, with reuse controls, tile styling and exact local export.

Free browser toolRuns in your browserNo sign-up

Your target and tile photos stay in this browser. This tool measures average tile colours and builds a real repeated-photo mosaic; it does not identify faces, generate images with AI or fetch unlicensed photos.

Target previewNo target

1. Add target photo

Choose target PNG, JPG or WebP

The target controls the overall composition. Up to 20 MB, 8192px per side and 64 megapixels.

2. Build tile library

Add multiple tile photos

2–100 tile photos, up to 10 MB each. More varied colours and subjects produce less repetition and better matching.

0 tile photos

3. Mosaic style

4. Grid and matching

Output long edgeTile columns50 × 50Tile diversity55%Match variation12%Variation seed7

Tile shapePhoto fit

Tile gap0 pxGap / tile background

5. Target and tile finish

Colour tint per tile18%Target image blend8%Target brightness100%Target contrast100%Target saturation100%

Tint and target blend improve long-distance recognition but reduce the visible independence of each tile. Set both to 0% for untinted tile photography.

Target preview

2000 × 2000 planned output

Preview shows the target until a mosaic is generated.

Generate mosaic

Planned output: 2000 × 2000, 50 × 50 cells.

Project and export

FilenameFormat

Quality92%

About the Photo Mosaic Maker

Photo Mosaic Maker creates an actual repeated-photo mosaic rather than placing unrelated pictures in a simple collage grid. It samples the target image cell by cell, measures the average RGB colour of every tile photo and selects a visually close tile for each position.

Upload one target and between two and 100 tile photos. Tile diversity penalizes repeated use, match variation introduces controlled alternatives among nearby colour matches and a deterministic seed creates another arrangement without making the result impossible to reproduce.

Choose 12 to 120 tile columns and a 1200, 2000, 3000 or 4000-pixel long edge. Square, rounded and circular tiles support crop-to-fill or contain fitting, optional gaps and a custom background colour while the target aspect ratio determines the final output dimensions.

Colour tint and target-image blend improve recognition from a distance, while zero-percent settings preserve unmodified tile colour more strongly. Target brightness, contrast and saturation are applied during colour sampling so the matching strategy can be tuned before generation.

Everything runs locally in the browser with generation progress and cancellation. Projects can embed the target and tile library for later regeneration, but the tool does not identify faces, remove backgrounds, generate imagery, fetch stock photos, guarantee every tile is used or replace professional large-format print preparation.

Key features

Real target-cell and tile-colour matching
One target PNG, JPG or WebP image
Bulk upload of up to 100 tile photos
Measured average RGB swatch for every tile
12 to 120 adjustable tile columns
Automatically calculated rows and preserved target ratio
1200, 2000, 3000 and 4000-pixel long-edge outputs
Square, rounded and circular tile shapes
Crop-to-fill and contain tile fitting
Adjustable tile gaps and background colour
Tile-diversity repetition penalty
Controlled alternative-match variation
Deterministic variation seed
Per-tile target-colour tint
Whole-target recognition blend
Target brightness adjustment for matching
Target contrast adjustment for matching
Target saturation adjustment for matching
Six practical mosaic presets
Visible generation progress
Cancellable long-running generation
Tile-library names, colour swatches and removal
Reported cell placements and unique tiles used
Undo and redo for up to 30 recent editing states
Versioned project save with embedded photographs
Strict project and image validation
PNG, JPG and WebP export with quality control
Responsive sticky preview for phone, tablet and desktop
Private local processing with no account or watermark

How to use

1Upload the target image whose overall composition the mosaic should reproduce.
2Add at least two tile photos; use a larger library with varied light, dark and colourful images for more accurate matching.
3Choose a balanced, detailed, people, circle, clean-grid or poster-distance preset as a starting point.
4Select the output long edge and tile-column count. More columns create smaller tiles and more cell placements.
5Adjust tile diversity to discourage excessive reuse and match variation to permit alternative near-colour photos.
6Choose square, rounded or circular tiles, crop or contain fitting, and optional gaps with a background colour.
7Tune colour tint and target blend while remembering that stronger values improve distance recognition but obscure independent tile colours.
8Adjust target brightness, contrast and saturation when the sampled composition needs lighter, stronger or less saturated matches.
9Generate the mosaic and review the reported grid, placement count and number of unique tile photos used.
10Download PNG, JPG or WebP, or save an editable project containing the source photos and settings for regeneration later.

Examples

Build a family portrait mosaic

Input: One licensed family portrait and 60 personal event photographs, Balanced Portrait preset

Output: 2000-pixel colour-matched mosaic preserving the portrait silhouette

Small private photographs become visible when viewers inspect the export closely, so obtain consent before publishing.

Create an anniversary number mosaic

Input: A high-contrast number graphic and 100 celebration photographs, Fine-Art Detail preset

Output: Dense 80-column mosaic with reduced repetition

High-contrast targets and broad tile-colour variety generally produce clearer shapes.

Make a circular social mosaic

Input: Square target, 30 travel photos, Circle Mosaic preset and dark gap colour

Output: Graphic dot-style mosaic with individually clipped photographs

Circular clipping removes tile corners; important subjects should be near the centre of each source photo.

Prepare a large event concept

Input: Landscape logo composition, 80 approved attendee images and 3000-pixel Poster Distance output

Output: High-density digital mosaic concept

For physical installation, confirm final dimensions, DPI, bleed, colour profile and photo permissions with the printer and organizer.

Frequently asked questions

Is this a real photo mosaic or just a collage?: It is a colour-matched mosaic. The target is sampled into cells and each cell receives a tile photo selected using measured colour distance plus the chosen diversity and variation controls.
How many tile photos can I use?: You can add two to 100 tile photos. Generation can repeat them across thousands of cells, and a more varied library normally produces better matches.
Will every uploaded photo appear in the mosaic?: Not necessarily. Selection is based on target colours, diversity and variation. A tile whose average colour is unlike every target region may remain unused; the result reports how many unique tiles were selected.
How do I reduce repeated photos?: Add more tile photos and increase Tile Diversity. Some reuse remains necessary whenever the mosaic has more cells than library images.
What do colour tint and target blend do?: Tile tint overlays each cell's target colour on its selected photo. Target blend overlays the overall target after tile placement. Both improve long-distance recognition but can make individual tile photography less natural.
Does the tool recognize faces or subjects?: No. Matching uses average colour, not facial recognition, object detection or semantic AI. Subject-aware placement is not claimed.
Are my photos uploaded to ToolsOnDuty?: No. Image analysis, matching, rendering, project generation and export are designed to happen locally in the browser.
Why can a 4000-pixel mosaic take longer?: Large outputs and dense grids require thousands of tile selections, clips and draws in browser memory. Generation yields periodically and includes a Cancel control, but speed still depends on the device and source images.
Does the project file contain my photographs?: Yes. The target and tile library are embedded so the mosaic can be regenerated later. Treat the project as sensitive and avoid sharing it when the source photos are private.
Is the export ready for large-format professional printing?: It provides the reported RGB pixel dimensions, not guaranteed physical dimensions, DPI metadata, CMYK conversion, bleed or a printer-approved proof. Confirm production requirements separately.
Can I publish a mosaic made from other people's photos?: Only when you have the required copyright, privacy, model-release and commercial permissions for the target and every tile image. A mosaic does not remove those rights.

Open a related tool to prepare your files or refine the finished result.