# Backend

The Rez backend is a modular Python-based system designed for high reliability and intelligent trade execution. It is divided into three primary functional engines:

## Components

### 1. [Trading Engine](https://docs.rezlabs.org/unit-architecture/backend/trading)

The gateway to the Hyperliquid exchange. It handles secure transaction signing, market data fetching, and order lifecycle management.

### 2. [Decision Agent](https://docs.rezlabs.org/unit-architecture/backend/agent)

The AI-driven core. It uses a ReAct pattern to orchestrate multi-turn reasoning cycles with Large Language Models, transforming indicators into trade decisions.

### 3. [Indicator Engine](https://docs.rezlabs.org/unit-architecture/backend/indicators)

The computational layer. It calculates technical indicators like EMA, RSI, and MACD locally using Binance data and `pandas-ta`, ensuring the agent has access to low-latency, accurate signals.

***

For a higher-level view, see the [System Overview](https://docs.rezlabs.org/unit-architecture/architecture).