Project Story

Inspiration

We were very inspired by the gap between fast-moving crypto markets and the tools everyday traders actually have. The idea for Tradini came from seeing how hard it is to turn noisy signals into clear decisions.

What it does

Tradini is an AI-assisted trading companion that explains market context, surfaces indicators, and helps users plan swaps with confidence. At a high level, it combines chat-based guidance with real-time crypto data and transaction-ready outputs.

How we built it

We built the frontend with React, TypeScript, and Vite for a fast, responsive UI. The backend uses Django + DRF for core APIs and an AI orchestration layer (CrewAI) to structure agent workflows, while a TypeScript service handles indicator calculations and market data. The architecture ties together authentication, data retrieval, and AI reasoning into one flow.

Challenges we ran into

We were very inspired but still hit challenges: stitching together multiple services, keeping responses fast while calling external APIs, and making AI outputs reliable enough for real trading decisions. Balancing explainability with concise guidance was another recurring challenge.

Accomplishments that we're proud of

We shipped a working, end-to-end system that combines AI chat, market indicators, and token operations in a single experience. We also integrated wallet-based auth and built reusable indicator endpoints that can scale with future features.

What we learned

We learned how to design a multi-service AI product that stays responsive under real-time data demands, and how to frame AI outputs with guardrails and transparency. We also deepened our understanding of trading math and indicators like VWAP:

What's next for Tradini

Next, we want to expand supported assets, add personalized strategies, and improve backtesting so users can validate ideas before acting. We also plan to refine the agent's risk-awareness and add clearer, user-controlled trade simulations.

Built With

• python