Insighter

Inspiration

We’ve always been curious about finance and investing, but quickly realized how hard it is to access professional-grade tools. Platforms like Bloomberg Terminal are powerful, but they cost more than most students, independent traders, or small teams could ever afford (upwards of $30k per year). We wanted to build something that improves accessibility to these kinds of insights without the barrier of cost. So, we have created Insighter.

What it does

Insighter is an AI-powered financial analysis dashboard that combines live market data with intelligent agents to guide users through the complexity of investing. It provides real-time stock quotes, historical charts, and market indices alongside news feeds enhanced with AI-powered summarization and sentiment analysis. Users can interact with specialized AI agents for research, portfolio advice, news analysis, and technical insights, while exploring stocks through a smart screener and interactive charting tools. The platform also integrates AI-driven diagram and image generation to visualize company relationships and market concepts, all presented in a flexible interface with floating windows for multitasking.

How we built it

Frontend: Next.js, React
Backend: FastAPI (Python)
API's: Polygon.io API (for financial data) and Google Gemini API with ADK (for gen-AI insights)
Infrastructure: Node.js

Challenges we ran into

Time Constraints: Building a full-stack platform with multiple AI agents, real-time data, and custom visualizations in just 36 hours pushed us to prioritize features while keeping stability.
Network Reliability: Running the project on campus eduroam Wi-Fi caused latency and dropped connections, making real-time data fetching more difficult.
API Quotas: Free-tier limits for Polygon.io and Gemini API quickly forced us to optimize calls, batch requests, and rethink our architecture to avoid hitting ceilings.
Integration Complexity: Getting multiple AI agents to operate in parallel without stepping on each other’s responses required careful orchestration.

Accomplishments that we're proud of

Successfully built a working AI-powered stock analysis platform within 36 hours.
Integrated real-time financial data with multiple AI-driven research agents.
Created a user-friendly dashboard that balances professional-grade insights with accessibility.
Learned how to coordinate frontend and backend integration efficiently under hackathon pressure.
Most importantly, we proved that a free, open alternative to expensive financial software is both possible and impactful.

What we learned

How to manage and optimize real-time data APIs within strict usage limits.
The importance of designing modular AI agents that can specialize but still collaborate.
How to use Google Gemini API effectively for summarization, analysis, and diagram generation.
How to rapidly prototype a full-stack application by splitting work efficiently across frontend, backend, and AI components.

What's next for Insighter

More personalization with customized dashboards and AI recommendations tailored to user portfolios.
Expanding AI agents by introducing deeper quantitative analysis with reinforcement learning agents.
Broaden Insighter as a teaching tool for students and beginner investors, increasing accessibility to financial education tools.