Indutus

Indutus turns photos of clothes into a weather-aware digital wardrobe, then uses Gemini to recommend and visualize real outfits for any occasion.

Comment

Indutus: AI-Native Personal Styling for the Wardrobe People Already Own

Indutus is an AI-native personal styling service that turns the clothes people already own into fast, practical outfit decisions. We are entering the Professional Services Access category because personalized styling is traditionally expensive, time-consuming, and inaccessible to most people. Indutus makes everyday styling guidance available through a digital wardrobe built from a user’s own clothing photos.

The problem is familiar: people may own many clothes but still spend time every day deciding what to wear. Fashion inspiration is often disconnected from a person’s actual wardrobe, local weather, occasion, body profile, and recent outfit choices. Indutus starts with what the user really owns. Users upload clothing photos, organize their wardrobe, choose an occasion such as work, travel, date, or casual wear, and receive a complete outfit recommendation with a generated visual preview.

We began building Indutus during the Build with Gemini XPRIZE submission period. The project was initialized on July 20, 2026. We use standard open-source web technologies, including React, TypeScript, Node.js, Express, MySQL, Docker, and Sharp, but the Indutus business workflow, AI orchestration, product experience, and deployment were created during this competition period.

Indutus is designed so that humans and AI have clear, complementary roles. Humans provide the source of truth: their actual garments, personal profile, city, preferences, and the occasion they are dressing for. Gemini performs the high-frequency work that would otherwise require manual cataloging and repeated styling research. It analyzes garment photos into structured wardrobe data, helps choose among viable outfit combinations, and generates a visual preview from the selected garment references. The user remains in control of what enters the wardrobe and whether to follow a recommendation.

In production, Gemini is not a decorative chatbot. It executes core decisions in the Indutus workflow. We use gemini-3.5-flash-lite for high-throughput multimodal garment understanding and outfit reasoning, and gemini-3.1-flash-lite-image to generate grounded outfit previews from the selected garments. The application uses the Gemini API in the live product experience.

The system also adds deterministic guardrails around the model. Before a recommendation is presented, Indutus checks outfit completeness, weather suitability, garment identity, layering, and recent-outfit repetition. The AI works from structured wardrobe records and selected-item references rather than being asked to invent a fashionable look from scratch. This is important because a useful styling recommendation must be wearable, appropriate for the situation, and based on clothes the user actually owns.

Speed is essential because outfit planning often happens immediately before people leave home. Across repeated end-to-end tests in the production environment, Indutus typically generates an outfit preview in about 20 seconds. Actual timing can vary with network conditions and upstream model response time, but the experience is designed to remain fast enough for interactive, everyday use.

Indutus has just launched and is currently in its initial promotion and customer-acquisition phase. We do not yet have paid customers. Our immediate business objective is to validate activation, retention, and willingness to pay through a subscription and credit-based model for high-frequency personal styling. We will report revenue, expenses, marketing spend, and customer evidence transparently in the required submission materials as the business develops.

The business opportunity is broader than simply generating outfit images. Indutus gives people access to personalized styling guidance that would otherwise require time with a professional stylist. By helping users understand and reuse the clothing they already own, it can reduce decision fatigue, help people prepare for work and important occasions, and make personal styling more affordable. Over time, the platform can also create opportunities for independent stylists, wardrobe consultants, and small fashion-service businesses to serve more clients using AI-assisted wardrobe organization and recommendation workflows.

What we learned is that useful AI styling depends on grounding. Generative output alone is not enough. The best results come from combining Gemini’s multimodal reasoning and image capabilities with structured garment data, clear user context, deterministic rules, and explicit validation. We learned that trust is earned when AI recommendations are fast, specific, wearable, and visibly connected to the user’s real wardrobe.

Indutus is live at https://indutus.com. We will provide product screenshots, production timing evidence, Gemini/API usage evidence, and launch-stage user feedback to demonstrate that AI is running in the product continuously. Our next goal is to turn early product usage into a repeatable business by improving onboarding, validating the paid model, collecting customer outcomes, and expanding the service from daily outfit selection into packing, wardrobe planning, and long-term personal style development.

Built With

Share this project:

Updates