Taurus

Algorithmic trading without technical barriers. Backtest & validate instantly

Comment

Inspiration

Taurus exists because automated trading shouldn’t be something only coders or big firms can use. I wanted a simple, friendly way for curious people to try algorithmic trading without writing code.

When you look at financial markets, it is just numbers at first: prices, timestamps, indicators. But behind those numbers is behavior.

Trends. Reactions. Signals.

Most people either rely on intuition or complex tools they don’t fully understand. We thought:

What if you could just describe your idea in plain English… and actually test it?

Instead of guessing whether a strategy works, we wanted to make it measurable, testable, and explainable.

What it does

Taurus is an AI-powered trading strategy builder that converts plain English ideas into executable backtests.

At its simplest, users can describe strategies like:

“Buy when a stock drops below a certain price”

Taurus translates that into structured logic, runs it against real historical market data, and outputs:

  • Performance vs the market
  • Alpha (performance)
  • Signal frequency
  • Time-based insights

Beyond just results, Taurus explains why a strategy worked by breaking it down over time showing trends, signal windows, and behavior after entry points.

How we built it

We designed Taurus as a modular system that separates data, logic, and analysis.

  • Python backend for strategy execution and backtesting
  • Alpaca API for real market data ingestion
  • Custom rule engine to translate structured strategy inputs into signals
  • Pandas for time-series processing and portfolio simulation
  • JSON-based inputs to allow flexible strategy definitions

We also built a time-analysis layer that allows us to:

  • slice specific date ranges
  • analyze post-signal performance
  • detect trends over time

This transforms raw price data into something comprehensible

Accomplishments that we're proud of

We built a fully functional backtesting pipeline from scratch.

Using Agentic Tool Calling for Google Gemini,

Taurus can:

  • take structured strategy inputs
  • fetch real market data
  • generate signals
  • simulate portfolio performance
  • output clean, analyzable results

We also created comparison outputs across multiple assets, allowing us to evaluate how a single strategy performs across different stocks.

Built With

Share this project:

Updates