DressRehearsal

Demo DressRehearsal

Inspiration

Fashion brands often make expensive production decisions before knowing what customers truly want. Traditional surveys ask people to imagine how a garment might look, while sales data arrives only after the collection has already been manufactured.

We saw an overlooked business opportunity: connecting potential customers directly with brands and manufacturers before production begins. Instead of asking whether someone likes a flat product image, DressRehearsal lets people see themselves wearing unreleased garments, choose their favorite, and indicate whether they would consider buying it at the intended retail price.

This creates an early feedback loop that can help manufacturers decide what clothing deserves to be produced, refined, or cancelled—before committing significant capital to inventory.

What it does

DressRehearsal is a virtual pre-production testing platform for fashion.

Customers first complete a short style profile covering occasions, silhouettes, colors, priorities, and preferred looks. They then upload a photo and use AI-powered virtual try-on to see themselves in three unreleased garments.

After viewing the results, each participant selects the garment they would most like to wear. Only once that choice is locked do we reveal the target retail price and ask whether they would consider buying it at that price.

This order is important: it separates emotional preference from commercial intent. A garment may receive the most votes but perform poorly once customers see its price. Another may attract fewer initial votes while generating stronger purchase intent.

Brands and manufacturers receive a private live dashboard showing:

The audience favorite based on personal preference.
The commercial favorite based on intent at the target price.
Style signals gathered during discovery.
Qualification rates for every garment.
Progress toward a predefined buyer-review threshold.
Cases where the most popular design and the strongest commercial opportunity are different.

The result is more useful than a simple popularity contest. DressRehearsal gives manufacturers evidence about what potential customers want them to make—and what they may actually buy—before inventory is produced.

How we built it

We built DressRehearsal with Next.js 16, React 19, TypeScript, and the YouCam AI Clothes API.

The customer experience guides participants through style discovery, consent, photo upload, virtual try-on, garment selection, and target-price feedback. YouCam generates the apparel previews from the participant’s photo and the reference images for each garment.

A server-side workflow manages uploads, creates three virtual try-on tasks, polls their progress, handles bounded retries, and returns the completed results without exposing provider credentials or signed URLs to the browser.

We used Node.js’s built-in SQLite support to store sessions, task states, style responses, preference votes, and purchase-intent signals. The brand dashboard aggregates this information into a live merchandising report that refreshes throughout the campaign.

Privacy was part of the architecture from the beginning. Uploaded photos are validated, stripped of metadata, normalized in memory, and never stored in our database or written to disk.

Challenges we ran into

One challenge was turning an impressive virtual try-on demo into a credible business tool. Generating an image is only the beginning; the product also needs to capture a decision that a manufacturer can use.

We designed the experience carefully to avoid price anchoring. Participants must choose their favorite garment before seeing its target price, allowing us to distinguish “I like this” from “I would consider buying this.”

Managing multiple asynchronous AI tasks was another challenge. Every session generates three separate results, which can complete or fail at different times. We implemented controlled concurrency, polling, timeouts, idempotent task creation, clear application states, and a limited retry flow.

We also had to protect sensitive user and provider data. Photos required strict validation and normalization, while API credentials, provider task IDs, upstream errors, and temporary result URLs had to remain server-side.

Finally, we worked to communicate the limits of the data honestly. A positive response is a signal of non-binding purchase intent—not a preorder, guaranteed sale, or complete demand forecast.

Accomplishments that we're proud of

We are proud that DressRehearsal connects a compelling consumer experience with a practical manufacturer decision.

For customers, it feels personal and enjoyable: they discover their style and see unreleased clothing on themselves. For brands, the same interaction produces structured pre-production evidence.

We are especially proud of separating audience preference from target-price intent. This helps expose an important commercial difference that traditional voting systems often miss.

We also created a live buying room where manufacturers can compare garments, identify emerging style patterns, and see which design has gathered enough evidence to deserve further review.

Most importantly, we built a product that could help reduce one of fashion’s most expensive problems: producing inventory based mainly on intuition and discovering too late that customers did not want it.

What we learned

We learned that virtual try-on can be more than an e-commerce conversion feature. Used earlier in the fashion lifecycle, it can become a research and merchandising tool.

Seeing yourself in a garment creates a much more meaningful response than looking at it on a model or answering a hypothetical survey. It makes early customer feedback more personal, concrete, and useful.

We also learned that popularity and commercial potential are not the same thing. Price must be introduced at the right moment to avoid distorting preference while still testing whether interest survives real-world constraints.

Finally, we learned that manufacturers do not necessarily need more data. They need clearer signals collected at the right stage of the production process.

What's next for DressRehearsal

The next step is to test DressRehearsal with real fashion brands and manufacturers during collection development.

We plan to support multiple brands, campaigns, product categories, and target customer panels. Manufacturers could upload early samples, invite relevant potential customers, compare designs, and collect feedback before approving production quantities.

Future reports could segment explicitly provided feedback by market, style profile, and customer group, helping brands understand not only which garment is preferred, but who wants it and in which context.

We also want to add structured qualitative feedback so participants can explain what they would change about a garment—such as its color, length, silhouette, material, or price.

Our long-term vision is for DressRehearsal to become a direct bridge between the people who make clothing and the people who may wear it: customers try future garments virtually, manufacturers learn what deserves to exist, and fewer unwanted products are made.