Inspiration: Local artisans and MSMEs frequently encounter significant friction when transitioning to digital commerce—underpricing erodes profitability, whereas overpricing deters prospective buyers. Simultaneously, online shoppers are constrained by rigid, fixed-price models with no structured mechanism to negotiate. We engineered Zuna to bridge this gap by creating an intelligent, flexible, and accessible marketplace tailored for small businesses.

What It Does: Zuna unifies four core modules into a cohesive commerce ecosystem:

  1. AI-Assisted Smart Pricing: Computes optimal listing prices using material, labour, packaging, and overhead inputs alongside inventory age, leveraging generative AI to deliver clear, human-readable pricing rationale.
  2. Dynamic Price Negotiation: Enables buyers to submit structured offers while empowering sellers to accept, reject, or counter within a protected minimum price floor.
  3. Creator-Driven Social Commerce: Connects sellers with content creators to promote listings through unique, performance-tracked referral links.
  4. Verified Razorpay Checkout: Locks mutually agreed prices server-side and routes transactions through Razorpay with cryptographic signature verification.

How We Built It: We developed a responsive frontend using HTML5, CSS3, and Vanilla JavaScript, supported by a RESTful Python and FastAPI backend with an SQLite relational database. We integrated a deterministic pricing engine with the Gemini API for contextual explanations and implemented Razorpay Standard Checkout (Test Mode) for secure payment processing.

Challenges & Key Learnings: A primary engineering challenge was preventing client-side price manipulation during post-negotiation checkout. Solving this reinforced the necessity of strict server-side trust boundaries—enforcing database-locked pricing and cryptographic payment verification—while demonstrating that transparent, explainable recommendations foster far greater seller confidence than opaque automation.

What's Next: Future iterations will introduce multilingual interfaces in Hindi and regional Indian languages, machine learning models trained on real transaction data for demand forecasting, and automated marketplace seller settlements.

Built With

Share this project:

Updates

Submission history