Dressly — Design Together, Create with AI

Inspiration

Fashion is one of the most personal forms of self-expression, yet the process of finding or creating the perfect accessory is often limited by predefined catalogs and static shopping experiences.

We wanted to explore a different approach: What if customers could collaborate with designers and artificial intelligence to bring their dream fashion products to life, visualize them instantly, and try them on virtually before making a decision?

That's how Dressly was born.

Our inspiration came from combining three powerful experiences: generative AI, virtual try-on technology, and real-time human collaboration. Instead of replacing human creativity, we wanted AI to enhance it, making personalized fashion design more accessible, interactive, and engaging.

What it does

Dressly is an AI-powered fashion customization platform that turns customer ideas into visual designs through generative AI, virtual try-on, and live collaborative consultations.

The platform brings together several capabilities:

🎨 AI-Powered Fashion Design

Customers can describe their ideal fashion accessories, including handbags, styles, colors, materials, and other visual preferences.

Using generative AI, Dressly transforms those descriptions into realistic product visualizations, allowing customers to explore different creative directions without requiring traditional design software.

👜 Virtual Try-On

By integrating YouCam's AI-powered virtual try-on technology, Dressly helps customers visualize fashion products in a more personal and immersive way.

This bridges the gap between imagining a design and understanding how it might look in real life.

🎥 Live Consultation — Design Together in Real Time

One of Dressly's most exciting features is Live Consultation, a collaborative design experience powered by Google Meet and AI.

Customers can join a video consultation with a designer to discuss their ideas, preferences, and desired changes.

But there's an additional participant: an AI-powered meeting assistant.

Our system uses a browser-based meeting bot with a human-in-the-loop approach to participate in Google Meet sessions and support the creative workflow.

During consultations, the AI assistant can interpret customer requests and design preferences from the conversation, helping translate natural language into actionable design suggestions.

For example, a customer might say:

"I love this handbag, but I'd prefer a darker shade of blue, gold details, and a more minimalist style."

Dressly can use those preferences to help generate an updated design through AI.

We also explored Picture-in-Picture (PiP) functionality to keep the video consultation accessible while customers interact with the design interface.

The goal is to make remote fashion consultations feel like sitting next to a designer and watching your ideas become reality.

✨ AI-Assisted Iteration

Customers can refine their designs through natural language, experimenting with multiple variations until they find something that matches their vision.

Rather than treating AI generation as a one-time action, Dressly makes it part of an iterative creative process.

How we built it

Dressly combines modern web technologies, generative AI, virtual try-on capabilities, and real-time communication.

Core technologies and integrations:

  • React: Powers the interactive frontend and fashion customization experience.
  • Runway API (Nano Banana 2.1): Generates and transforms fashion product visuals based on customer descriptions and preferences.
  • YouCam AI Try-On: Provides virtual try-on capabilities to help customers visualize products.
  • Google Meet: Enables real-time video consultations between customers and designers.
  • Browser automation with Chrome: Supports the meeting assistant that participates in Live Consultation sessions.
  • Human-in-the-loop architecture: Keeps people involved in the meeting assistant workflow.
  • Picture-in-Picture: Allows video consultations to coexist with the interactive design experience.

A key architectural decision was to build Dressly as an integrated experience rather than a collection of disconnected AI tools.

The challenge was not simply generating images or joining video calls. It was making the entire process feel natural: from a customer's initial idea to AI-assisted visualization, virtual try-on, collaborative feedback, and design refinement.

Challenges we ran into

1. Bridging human conversation and generative AI

One of the biggest technical challenges was connecting the natural flow of a video consultation with a structured AI design workflow.

People don't communicate design requirements as perfectly formatted prompts. They change their minds, reference previous ideas, and describe things subjectively.

We needed an approach capable of interpreting these conversations and turning them into useful design instructions without disrupting the creative process.

2. Integrating an AI assistant into Google Meet

Google Meet was designed primarily for human communication, not autonomous AI participants.

Building a browser-based assistant required exploring Chrome automation, meeting participation, permission handling, and human-in-the-loop controls.

Keeping the interaction practical and manageable while respecting the consultation experience was an important design challenge.

3. Maintaining visual consistency

Generating beautiful images is relatively straightforward with modern generative models. Maintaining the identity of a specific fashion product across multiple design iterations is much harder.

Changing a handbag's color or material should not unintentionally alter its overall shape or defining characteristics

Built With

Share this project:

Updates

Submission history