Vibe Coding Hackathon 2026 - Winners Announced
The results are in. After an intense round of AI-assisted building, HackIndia is proud to announce the winners of the Vibe Coding Hackathon 2026, powered by Sharp Economy.
Every submission was evaluated against a rigorous, evidence-based 100-point rubric covering depth of AI integration, real-world problem solving, deployability, code quality, complexity, usability, and depth of knowledge. Repositories were cloned and analyzed for code volume, AI provider wiring, frameworks, live deployment configuration, and demo quality. Placeholder URLs were discarded. Only genuine, working builds made the cut.
Here are the top 3 teams who built the best of the best.
🥇 1st Place — Kasukabe-Coders
MandiMind AI — 91/100
Described as the "Bloomberg Terminal for Indian farmers," MandiMind AI is built directly on government open mandi data, giving farmers real-time, intelligent market insights in a single live platform. The team shipped a working live demo along with a Loom walkthrough video, and integrated multiple AI models including Claude, Groq, and Llama into the build.
Score Breakdown
Vibe/AI Build Use: 20/20 · Problem Solving: 20/20 · Usability: 10/12 · Deployability: 11/14 · Code Quality: 14/14 · Complexity: 8/12 · Depth of Knowledge: 6/8
🥈 2nd Place — Arclight
CogniFlow — 89/100
CogniFlow is a full-fledged AI analytics dashboard built on React, Express, and Vite, integrating a broad AI stack spanning Anthropic, Gemini, Groq, and LangChain. The team deployed cleanly to Vercel with strong, production-ready code quality throughout.
Score Breakdown
Vibe/AI Build Use: 20/20 · Problem Solving: 18/20 · Usability: 10/12 · Deployability: 14/14 · Code Quality: 14/14 · Complexity: 7/12 · Depth of Knowledge: 6/8
🥉 3rd Place — Driftx
RoadSense AI — 87/100
RoadSense AI tackles road safety head-on with an accident-risk platform for India, featuring voice alerts and automated PWD repair notices for damaged roads. The team built a full-stack deployment with frontend on Vercel and backend on Railway, scoring a perfect 12/12 on usability.
Score Breakdown
Vibe/AI Build Use: 19/20 · Problem Solving: 18/20 · Usability: 12/12 · Deployability: 14/14 · Code Quality: 11/14 · Complexity: 7/12 · Depth of Knowledge: 6/8
How Teams Were Judged
Every submission was scored against a transparent, weighted rubric designed to reward genuine AI-assisted building, not just surface-level demos.
"Vibe / AI Build Use" measured the depth of AI integration and evidence of an AI-assisted, vibe-coding workflow using tools like Cursor, Lovable, Copilot, Windsurf, or Tempo.
Each repository was independently cloned and evaluated for code volume, AI provider wiring, frameworks used, deployment configuration across platforms like Vercel, Netlify, Render, Railway, Docker, and Fly, genuine live URLs, README substance, and demo video quality. A small bonus was awarded for working demo videos. Placeholder or non-functional URLs were excluded entirely from deployability scoring, keeping the evaluation grounded strictly in evidence.
Prizes
🥇 1st Place: 10,000 SHARP tokens
🥈 2nd Place: 5,000 SHARP tokens
🥉 3rd Place: 2,000 SHARP tokens
Prizes awarded courtesy of our hackathon sponsor, Sharp Economy.
Thank You
Congratulations to Kasukabe-Coders, Arclight, and Driftx for setting the bar at Vibe Coding Hackathon 2026. A huge thank you to every single participant who showed up, built, shipped, and pushed the limits of what AI-assisted development can do.
Special thanks to Sharp Economy for powering the prize pool and supporting India's vibe coding community.
This is what happens when builders are given the tools, the time, and the right push. See you at the next one.