Outdrobe

Landing Page
Kloset
Wardrobe
Marketplace
Landing (Light)

Inspiration

After brainstorming for two weeks straight, we hit a creative dead end. Like a writer's block they say. That was until we looked back at our journey as international students arriving at college. The culture here was so different, and even something as simple as what to wear to class felt confusing. We realized that how you present yourself deeply affects your confidence, but with a suitcase full of clothes and no clear combinations, figuring out what to wear became its own daily puzzle.

That’s when the idea for OutDrobe was born — a personalized, voice-integrated agentic AI stylist that scans your own collection, creates unique outfit combinations, and even helps you shop for missing pieces by finding the cheapest matching options online.

When we looked around, we couldn’t find any all-in-one fashion solution — some apps let you upload your wardrobe, others just linked products. We wanted something that does it all: an assistant that plans your looks for classes, interviews, parties, or even social events like Halloween. (Trust us — the year we ended up wearing just a sticker to a Halloween party was the last straw 😅.)

OutDrobe ensures you never end up in that situation again.

What It Does

OutDrobe is your AI-powered fashion companion — an intelligent stylist that lives inside your closet. Unlike traditional styling apps or retail recommendation engines, Outdrobe is not a shopping catalog—it is a design intelligence system built around personalization and context. While competitors focus on selling products or static outfit templates, Outdrobe starts with what users already own and builds upward. It uniquely combines design logic, personalization, and real-time contextual understanding to create outfits that are both expressive and functional. By bridging the gap between personal wardrobe data and the global fashion ecosystem, Outdrobe offers an experience closer to having a personal stylist — but infinitely faster, smarter, and more accessible. Outdrobe transforms clothing choice into a science of confidence — redefining how people dress, travel, and express themselves through style. It: 🧠 Analyzes your wardrobe with a quick snap to identify item types, colors, and categories.

🪄 Creates new outfit combinations for any event, season, or vibe.

💬 Speaks to you — thanks to integrated voice AI — so you can have a natural conversation about what to wear.

🧠 Remembers every outfit you’ve worn, avoiding repetition and helping you maintain fresh looks.

🛍️ Shops for you — fetching affordable, matching pieces if your wardrobe is missing something.

💅 Learns your preferences over time, understanding your personal style and improving with every use.

Whether it’s a presentation, a coffee date, or a costume party, OutDrobe ensures your outfit fits the moment — perfectly.

How We Built It

Here is the tech stack we used: Frontend: Figma for the mockup and initial draft. Later built using TypeScript, Vite, React.js, and Three.js for a fast, interactive experience. Designed a minimal and aesthetic wardrobe dashboard for users to upload and visualize their collections.

Backend:Developed in Python (Flask + PyTorch) with Firebase for data storage. Integrated Google Cloud APIs for vision tagging and hosting services. Used BLIP for image captioning and feature extraction from wardrobe photos.

AI Agent: Built using Elastic Agentic AI for reasoning and autonomy — allowing the system to plan outfits, detect missing items, and take actions autonomously.

Voice Integration:Used Fish Audio for real-time speech-to-text and text-to-speech. Enabled full conversational interaction — users can talk directly to their AI stylist.

Challenges We Ran Into

First time building an agentic AI — integrating reasoning, memory, and actions took a lot of experimentation. Voice integration was a steep learning curve — balancing latency, response flow, and conversation quality that gives responses according to user's mood. Diving deep into Elastic’s documentation and wiring it all together took time (and more Red Bulls than we’d like to admit) but we pulled through!

Accomplishments We’re Proud Of

We built something we genuinely want to use every day. Successfully connected voice, vision, and agentic reasoning in one system — that’s rare for hackathon projects. Made the AI fun, witty, and personalized, giving it a real personality. And last but not the least, delivered a functional, interactive demo with amazing UI/UX to make it more user friendly ! (Came from our frustration from seeing Powerpoint Websites around the web -sighs. )

What We Learned

We learned how to combine multiple AI domains — vision, reasoning, and voice — into one cohesive experience. Appreciate the power of agentic frameworks like Elastic, and the importance of persistent memory in user-facing AI. Most importantly, we learned that good tech is only impactful when it’s relatable — and that solving our own daily frustrations often leads to the best ideas.

🚀 What’s Next for OutDrobe

We plan to turn OutDrobe into a fully-fledged startup — a well-rounded AI stylist platform that personalizes fashion for real users. Future goals include: -Integrating AR-based virtual try-ons. -Expanding wardrobe understanding through body type and fabric awareness. -Building a social layer, where users can share fits or compete in daily fashion challenges.

OutDrobe is just getting started ! The future of AI-powered personal styling is already hanging in your closet and is just one call away.