nexoraa - next genaration exploration or ui-navigator-agent

nexorra _ next gen ui navigator agent

Inspiration

Every Indian wastes 45+ minutes daily navigating MakeMyTrip, IRCTC, Swiggy, and government portals. A human concierge costs ₹25,000/year — unaffordable for 95% of Indians. We asked: what if AI could do it all for free?

What it does

Nexoraa is an AI agent that sees your screen and completes digital tasks automatically. Just say "Book me a flight to Delhi" — it navigates the app, fills the form, and confirms before payment. Handles travel, food, bills, government schemes, and more.

How we built it

Built using Llama 3.3 70B via Groq for intent parsing, Gemini Vision for screen understanding, Node.js backend with Express, and Playwright for browser automation. Deployed on Render with GitHub CI/CD.

Challenges we ran into

Making AI reliably understand dynamic UIs, handling CAPTCHAs, and building safe confirmation gates before payments without breaking the conversation flow.

Accomplishments that we're proud of

A fully working AI agent that navigates real websites end-to-end — not just a chatbot, but one that actually acts.

What we learned

The hardest part isn't the AI — it's reliable execution. Small prompt changes produce dramatically different results. Speed matters enormously in agentic UX.

What's next for Nexoraa

WhatsApp integration for rural users, voice-first mode in 8 Indian languages, real payment gateway support, and a full multimodal screen navigator powered by Gemini 2.0 — making every app accessible to every Indian