The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →
ThatPainter is reader-supported. When you buy through links on our site, we may earn an affiliate commission. Learn More
Google’s “latest AI experiment” was actually Runway Palette, a fashion-archive project announced in 2019. Built with The Business of Fashion, it used machine learning to organize runway looks by color, allowing visitors to discover designers and collections through visual similarity instead of starting with a text search.
Runway Palette was Google’s 2019 color-search experiment for fashion
Google’s “latest AI experiment” was not a newly launched 2026 fashion service. It was Runway Palette, an experiment announced on November 22, 2019 by Google Arts & Culture in partnership with The Business of Fashion (BoF).
Its unusual idea was simple but powerful: instead of beginning with a designer’s name, a brand, or a garment category, users could explore runway fashion through color. Machine learning analyzed the dominant colors in runway photographs and organized the archive into an interactive visual index.
Runway Palette was therefore an archive-discovery tool—not a virtual fitting room, clothing generator, shopping engine, or automated stylist. Its importance was showing how AI could create a new way to search a large cultural collection.
#1 Best Overall
What was Runway Palette?
Runway Palette turned BoF’s runway-photo collection into a color-based fashion visualization. According to Google’s announcement, the collection contained approximately 140,000 photographs from almost 4,000 fashion shows.
Google said the project represented approximately four years of fashion and nearly 1,000 designers. These are Google’s reported project-scale figures, rather than independently audited totals.
The system extracted the main colors from each runway look and used machine learning to organize the photographs according to their palettes. That allowed visitors to move through relationships between colors and discover looks that shared a similar visual character.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallCrashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteThe experiment was one of two projects Google presented as examples of AI being used in creative fields. The other, called Living Archive, explored AI-assisted dance choreography.
How the AI worked
The documented process had two important stages:
- Color extraction: The system identified the dominant colors in runway images.
- Machine-learning organization: It arranged the images by color palette so that users could browse related looks.
Google did not publish the model architecture, training code, clustering method, accuracy measurements, or a formal evaluation of the system in the announcement. It would therefore be inaccurate to describe Runway Palette as using a particular neural-network architecture or to claim that its color matches had a measured accuracy rate.
The important point is what the system made possible: a visual property of an image—its color composition—became a navigation layer over a fashion archive.
How people explored the fashion archive
Runway Palette combined three kinds of interaction.
1. Browsing by palette
Users could explore a visual color space and move toward runway looks with related combinations. This reversed the usual order of fashion search. A visitor did not need to know the name of a designer or collection before beginning; a palette could be the starting point.
2. Filtering the results
The experience allowed users to examine the archive through categories including:
- color;
- designer;
- season; and
- trend.
Google described the archive as covering fashion weeks from around the world. The filters connected the visual browsing experience to conventional fashion metadata, allowing a user to move from a color relationship to the designers and seasons associated with it.
3. Comparing a personal photograph
Visitors could also take or upload an image—for example, clothing from their own wardrobe or a photograph of autumn leaves—and compare its palette with runway looks.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
This did not mean that Runway Palette identified an exact purchasable garment. Its documented function was to find visually similar color combinations within the runway archive. A photograph supplied the visual starting point; the results led back to fashion imagery rather than to a product catalog.
Why using color as a search interface mattered
Fashion archives are normally organized around words: designer names, collection titles, seasons, garments, locations, and dates. Those labels are useful, but they require the visitor to know what to ask for.
Color provides a different entry point. Someone might see a deep blue-and-orange combination in a photograph, notice the colors in a wardrobe, or simply want to investigate a mood without knowing its fashion vocabulary. Runway Palette translated that visual impression into a way of navigating the archive.
That makes the project an example of AI-assisted indexing and discovery. The machine-learning system did not replace the archive with newly generated images. Instead, it analyzed existing material and helped visitors find connections that would be difficult to locate through text alone.
For a painting audience, the concept is especially familiar: a viewer can respond to hue, value, saturation, and color relationships before knowing the artist or title. Runway Palette applied a related idea to runway photography, using palette similarity as a bridge between personal observation and a large historical collection.
What Runway Palette did not do
Several descriptions of fashion AI can sound interchangeable, but Runway Palette belonged to a more specific category. It did not:
- generate original clothing designs;
- place garments onto a photograph of the user;
- act as a virtual fitting room;
- recommend exact products for purchase;
- predict fashion trends with a documented accuracy rate; or
- replace a stylist or provide a complete personal wardrobe plan.
Google’s announcement included trend browsing, but that is not the same as proving that the system forecast future trends. The safest description is that Runway Palette let people explore trend-related runway material within the archive.
Rank #4
Runway Palette versus Google’s later fashion tools
Google has since presented other visual and fashion-related experiments, but they should not be folded into the 2019 project.
The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →| Tool or project | Primary purpose | How it differs from Runway Palette |
|---|---|---|
| Runway Palette | Explore a runway archive through color relationships | Focused on cultural-archive discovery and runway imagery |
| Color Palette / Art Palette | Use colors from a photograph to find visually related artworks and cultural objects | Related color-search concept, but centered on art and cultural collections rather than the BoF runway archive |
| Google Shopping virtual try-on | Let users upload a photo and visualize clothing on themselves | Shopping and product visualization, not historical runway exploration |
| Doppl | Explore outfits and visualize personal style | A later Google Labs fashion experiment aimed at personal styling |
| Circle to Search fashion tools | Search for fashion items and related products from visual references on supported devices | Visual shopping and product discovery rather than color-indexing an archive |
These later initiatives show a broader movement from cultural-collection exploration toward personal styling, visual product search, and virtual try-on. They are not evidence that Runway Palette itself offered those capabilities.
Can you still use Runway Palette?
Google originally said that its experiments were available through the Google Arts & Culture experiments page and the service’s free iOS and Android apps.
Google Arts & Culture’s current materials continue to reference Runway Palette as a way to search The Business of Fashion’s catwalk archives. However, the available evidence does not establish that the original standalone experience works identically in every country, browser, or app version.
There is also a current Google Arts & Culture feature called Color Palette, sometimes described as Art Palette. It uses colors from a user’s photograph to find artworks and other cultural objects with related palettes. That feature is not the same as Runway Palette. A current app listing also describes fashion content alongside the wider Google Arts & Culture collection, which spans institutions and cultural material around the world.
Recommended Free Tools
In practical terms, the safest wording is:
- Google launched Runway Palette in 2019.
- The experiment allowed color-based exploration of BoF runway imagery.
- Google Arts & Culture still references the project in current fashion materials.
- The original interface should not be promised as universally available or fully functional for every user today.
The project’s lasting significance
Runway Palette was an early, clearly defined example of machine learning being used to make an image archive more discoverable. Its innovation was not that AI produced a new dress or automatically sold one. It was that AI helped reorganize existing visual material around a question people might not otherwise be able to express in words:
Show me fashion that feels like this color combination.
That distinction still matters. Generative fashion systems create or modify images. Shopping systems identify products and support transactions. Virtual try-on systems simulate garments on a person. Runway Palette instead connected a viewer’s visual input to a curated runway archive.
For artists, designers, and anyone studying color, the experiment offers a useful model for thinking about digital collections. A large archive can be searched not only by authorship and chronology but also by visual qualities—palette, composition, texture, or form—provided those qualities can be extracted and represented meaningfully.
Frequently Asked Questions
What was Google Runway Palette?
Runway Palette was a Google Arts & Culture experiment announced in November 2019 with The Business of Fashion. It organized runway photographs by their dominant color palettes so visitors could explore designers, seasons, trends, and visually related looks.
How large was the Runway Palette archive?
Google reported that the archive contained approximately 140,000 photographs from almost 4,000 fashion shows, representing about four years of fashion and nearly 1,000 designers. Those figures are reported project-scale totals, not an independently audited count.
Did Runway Palette generate clothes or let users try them on?
No. Runway Palette was designed for visual discovery within a runway archive. It did not generate clothing, simulate garments on a person, identify guaranteed purchasable products, or function as a virtual fitting room.
Is Runway Palette still available?
Google Arts & Culture still references Runway Palette in current fashion materials, and its app includes a separate Color Palette feature for finding artworks with colors related to a user’s photograph. The available evidence does not guarantee that the original Runway Palette interface is fully functional for every region, browser, or app version.
Free tools Windows power users keep installed
One-click scans. No signup required.
The Bottom Line
Runway Palette was Google’s 2019 experiment for exploring The Business of Fashion’s runway archive through color. It used machine learning to extract dominant colors from approximately 140,000 runway photographs and helped users browse related looks by palette, designer, season, and trend. Users could also compare a personal photograph with runway color combinations. It was an archive-discovery tool—not a clothing generator, virtual fitting room, trend-prediction system, or guaranteed shopping service.
Quick Recap
