Inspiration
Home renovation can quickly descend into absolute chaos. A few years ago, when my family bought a new home, we anticipated a smooth, exciting transition. Instead, it became a four-year living nightmare. We trapped ourselves in an endless cycle of low-quality workers, predatory pricing, and hidden scams where contractors constantly pushed unqualitative, overpriced materials. At our lowest point, a targeted robbery occurred: our premium, expensive Italian water heater and radiators—the only valuable assets in the house at the moment—were stolen. While the police investigation yielded nothing, it was clear an inside worker was responsible.My story is far from an exception. Globally, the systemic friction of independent contracting can be modeled by a high probability of project distress. If we define the probability of an average homeowner encountering severe budget, timeline, or security overruns as P(D), statistical market trends reveal: P(D) = 0,50. This means exactly every second person faces a catastrophic renovation bottleneck. I created this AI Home Renovation Assistant so that future homeowners have a digital shield, ensuring their structural journeys are safe, transparent, and significantly less horrible.
What it does
The platform acts as an intelligent coordinator designed to protect the homeowner's capital and peace of mind through seven core characteristics: 1) Objective Diagnostics: Instantly identifies the client’s structural goals and raw budget limits. 2) Market Demystification: Educates the client on localized sector pricing so they understand how the segment operates. 3) Dynamic Mapping: Stretches out an interactive blueprint map for the house matched against real-time budget scaling. 4) Global Price Aggregation: Scrapes and compares material prices, structural quality, and delivery duration across the web. 5) Vetted Specialist Matching: Sources top-tier specialists by cross-referencing past client feedback and verified assessments. 6) Scam Prevention & Legal Guardrails: Automatically flags predatory quotes and writes out sound legal standings/agreements for working specialists. 7) Design Ideation: Assists with foundational layout aesthetics or matches the user with professional designers. While standard e-commerce comparisons are simple, local raw-material hubs (like the chaotic Eliava-type physical markets) lack structured APIs. To solve this, the AI is designed to dynamically parse regional Facebook and Instagram pages, actively messaging vendor accounts to negotiate pricing, verify material quality, and audit past client feedback directly.
How we built it
The core architecture was built utilizing the ASI:ONE framework. The platform orchestrated the underlying multi-agent setup, executing the heavy lifting of agent creation, communication loops, and background deployment logic. To keep the project aligned with real-world contextual data and fine-tune prompt strategies, I closely assisted the workflow using Gemini to refine the interface logic. Final deployment happened on Agentverse platform with help of ASI:ONE once again.
Challenges we ran into
Everything collided with my drive to deliver a perfectly styled, complete web application. Our first major bottleneck appeared when the automated ASI:ONE link on the Hands-On Lab (HOL) interface encountered runtime errors, causing the core chat window to freeze.
From there, the real coding battle began. I pushed through successive local and cloud environment deployments—testing configurations across Vercel, Netlify, and GitHub—trying to force a complex multi-agent layout to cooperate with standard web hosting. Combined with persistent connectivity issues within Agentverse, achieving a seamlessly unified layout and live chat window proved to be an uphill battle against the clock.
Accomplishments that we're proud of
Despite every technical wall, framework error, and deployment blocker thrown my way, I managed to build a working conversational chat interface that successfully processes and answers user prompts. Seeing the system successfully output responses after hours of troubleshooting was incredibly rewarding.
What we learned
This was my very first hackathon, and while I entered the competition wanting every single deployment piece to be absolutely flawless, reality taught me a vital engineering lesson. I truly learned the meaning behind the classic industry motto: "Make it work first; make it perfect later." Pushing through the technical chaos gave me invaluable, firsthand experience in rapid prototyping under immense pressure.
What's next for Untitled
The immediate next step is to iron out the deployment configurations across our hosting stack to take the user interface fully live. Beyond the infrastructure fixes, we plan to expand the agent network's natural language processing capabilities, allowing the AI to smoothly automate text-based inquiries to local social media suppliers and aggregate non-traditional market data seamlessly.
Log in or sign up for Devpost to join the conversation.