Pivotal.ai is a full-stack, data-driven application that utilizes a custom-built Agentic AI system to identify high-probability swing trading opportunities in the stock and options markets.
This project is designed to showcase mastery in building secure, highly scalable, and disciplined full-stack applications, adhering to best practices like Test-Driven Development (TDD) and Clean Architecture.
Demo Link: https://pivotal-ai-web-app.vercel.app/
Warning
Currently Under Construction - Video Last Updated @ 1/11/2026 ~ In Development
In the chaotic world of trading data, Pivotal.ai acts as an intelligent scout. It replaces manual analysis by processing multiple market indicators and translating complex data into a concise, actionable trading recommendation (BUY/SHORT) complete with a target price and risk assessment.
- Indicators Processed: Price Action, Relative Strength Index (RSI), Moving Averages (e.g., 50-day SMA).
- Output: Actionable trade recommendation, Target Price, Stop-Loss/Risk Assessment.
This application leverages a modern, decoupled stack for security, speed, and maintainability.
| Component | Technology | Rationale |
|---|---|---|
| Backend / API | Django (Python), Django REST Framework | Chosen for robust transactional integrity, built-in security features, and the mature ORM integration with PostgreSQL—critical for a financial application. |
| Agentic AI Logic | Gemini API (Python), Custom Tools | The core intelligence. The agent is prompted with market data and defined trading rules, operating independently to generate actionable insights. |
| Data Source | Alpha Vantage API (or similar) | Provides reliable, granular historical and real-time data necessary for calculating technical indicators. |
| Frontend / UI | Next.js (React) | Provides a fast, modern, and SEO-friendly user interface, capable of displaying interactive charts and real-time trade alerts. |
| Database | PostgreSQL | Utilized for its superior transactional reliability and advanced indexing capabilities required for storing financial data securely. |
This project emphasizes financial rigor and software discipline through several key architectural choices:
-
All critical business logic, especially in the Django services (calculating indicators, running agent prompts), was built using a Test-Driven Development (TDD) approach to ensure near 100% Branch Coverage for all financial calculations.
-
All simulated debits/credits and balance updates utilize
django.db.transaction.atomic()to guarantee ACID compliance (Atomicity, Consistency, Isolation, Durability) and prevent data corruption in the trading log. -
The Python agent logic demonstrates the ability to invoke external tools (the Alpha Vantage data fetcher) and synthesize that information based on a rigorous, system instruction prompt, showcasing sophisticated LLM utilization.
-
The Agent logic, the data fetching, and the API request handling are separated into distinct service layers, maximizing code clarity, maintainability, and testability.