Skip to content

Latest commit

 

History

212 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Quantt

Local-first algorithmic cryptocurrency trading platform focused on execution, analytics, and extensibility.

Status Version License Python Node


Overview

Quantt is a modular cryptocurrency trading platform designed for local execution, strategy experimentation, and advanced market analysis.

The project combines:

  • Multi-exchange trading via CCXT
  • Technical analysis pipelines
  • AI-assisted forecasting
  • Custom risk management
  • Backtesting and simulation tooling
  • Local-first architecture with optional future cloud expansion

Quantt prioritizes lightweight execution, transparency, and extensibility while maintaining direct user control over trading operations.


Goals

Primary Objectives

  • Automate cryptocurrency trading workflows
  • Improve operational consistency and execution speed
  • Provide customizable strategy development
  • Support quantitative experimentation and forecasting
  • Maintain local-first execution and data ownership

Non-Goals

Quantt is not intended to:

  • Replace user trading decisions entirely
  • Provide financial advice
  • Manage user finances beyond broker interaction
  • Operate as a custodial platform

Target Users

  • Algorithmic trading enthusiasts
  • Quantitative traders
  • Developers experimenting with market automation
  • Companies evaluating engineering or quantitative talent

Current Development Roadmap

Version Milestone Status
v0.1 Execution engine (local-only) Complete
v0.2 Backtesting integration Planned
v0.3 AI forecasting layer Planned
v0.4 Optional cloud backend Experimental / Possible
v1.0 Stable public release Future

Core Features

Trading Engine

  • Real-time order execution
  • Exchange integration through CCXT
  • Order lifecycle management
  • Market condition monitoring
  • Custom strategy execution

Analytics

  • Technical indicators
  • Market signal generation
  • Performance metrics
  • Graphical analytics
  • Trade evaluation

AI Forecasting

  • Time-series forecasting models
  • OHLCV-based predictions
  • Indicator-aware inference
  • Lightweight inference pipeline
  • CPU-friendly execution

Risk Management

  • Configurable risk parameters
  • Position sizing logic
  • Exposure limitation
  • Strategy-specific safeguards

Local-First Architecture

  • Fully local execution
  • Local API key storage
  • No custody of user funds
  • Reduced external dependencies

Supported Exchanges

Current and planned integrations through CCXT:

  • Binance
  • Bybit
  • OKX
  • MEXC

Tech Stack

Languages

  • Python
  • TypeScript
  • JavaScript

Frontend

  • React
  • Vite
  • Electron

Backend / Engine

  • FastAPI
  • SQLAlchemy
  • CCXT
  • loguru

Quantitative & AI Stack

  • PyTorch
  • vectorbt
  • Pandas
  • NumPy
  • empyrical

Architecture

┌─────────────────────────────┐
│  Electron + React Frontend │
└──────────────┬──────────────┘
               │
               ▼
┌─────────────────────────────┐
│          FastAPI           │
└──────────────┬──────────────┘
               │
               ▼
┌─────────────────────────────┐
│    Trading Engine Core     │
│      (OOP Structured)      │
└──────────────┬──────────────┘
               │
               ▼
┌─────────────────────────────┐
│     CCXT Exchange Layer    │
└──────────────┬──────────────┘
               │
               ▼
┌─────────────────────────────┐
│          Database          │
└─────────────────────────────┘

AI Forecasting Layer

Model

  • Granite TTM 2.1 (fine-tuned)

Forecast Inputs

  • OHLCV market data
  • Active technical indicators
  • Market condition metrics

Precision Modes

  • FP64
  • FP32

Performance Goals

  • Lightweight execution
  • Minimal CPU overhead
  • Usable on low-end systems
  • Designed for 2-core CPUs

Installation

Requirements

Component Requirement
CPU 2 cores minimum
RAM 2 GB recommended
Storage ~200 MB
Python 3.11+
Node.js 25+
OS Linux / Windows / macOS

Quick Start

Clone Repository

git clone <repo>
cd quantt

Engine Setup

cd quantt-engine

py -m venv venv

# Linux / macOS
source venv/bin/activate

# Windows
venv\Scripts\activate

pip install .

py main.py

UI Setup

cd quantt-ui

npm install

npm run electron
npm run vite

Environment Variables

API_BINANCE=
API_SECRET_BINANCE=

API_BYBIT=
API_SECRET_BYBIT=

API_OKX=
API_SECRET_OKX=

API_MEXC=
API_SECRET_MEXC=

Configuration

Configuration Directory

Engine/config

Notes

  • Contains exchange-related settings
  • Stores local configuration values
  • May contain sensitive API-related information

Usage

Input

User-defined strategy and execution configurations via UI.

Output

  • Performance metrics
  • Strategy analytics
  • Risk information
  • Trading statistics
  • Graphical reports

API

Default Ports

Service Port
Engine 8000
UI 5173

Endpoints

Defined in:

Engine/main.py
Ui/

Trading Modes

Backtesting

  • Executed via UI
  • Powered by vectorbt
  • Historical strategy evaluation

Paper Trading

  • Demo-mode execution

  • No real funds involved

  • Supported exchanges:

    • Binance
    • Bybit

Live Trading

Requires:

  • Exchange API credentials
  • Proper configuration
  • User responsibility

Project Structure

quantt/
├── quantt-engine/
│   ├── Engine/
│   ├── config/
│   ├── qdata/
│   ├── logs/
│   └── reporting_portifolio/
│
├── quantt-ui/
│   ├── src/
│   ├── public/
│   └── electron/
│
└── README.md

Security

Important Notes

  • API keys are stored locally
  • Quantt does not custody user funds
  • Users are responsible for securing credentials
  • No trading occurs without proper API configuration

Safety Defaults

  • Demo mode recommended by default
  • Explicit API configuration required
  • Local-first execution model

Limitations

  • Primarily designed for local execution
  • Limited horizontal scalability
  • Single-exchange session support (current)
  • AI forecasting still experimental

Contributing

Contribution Rules

  • Contributions only through GitHub pull requests
  • Large architectural changes may have lower acceptance probability
  • Ensure proper testing before submission

Code Style

Naming Conventions

snake_case
SCREAMING_SNAKE_CASE

Formatting

  • 4-space indentation
  • Structured OOP organization
  • Modular component separation

Screenshots

Home

Home Screen

Resume / Analytics

Resume Screen


Disclaimer

Quantt is an experimental trading platform.

This software does not provide financial advice. Cryptocurrency trading involves substantial risk, including the possibility of capital loss.

Users are solely responsible for:

  • Trading decisions
  • API credential management
  • Risk management
  • Financial outcomes resulting from usage

License

MIT License


Maintainer

  • THI100 (solo project)

About

A trading algorithm for Crypto markets, automated and custom methods, utilizing CCXT for multi-plataform usage.

Topics

Resources

Stars

2 stars

Watchers

1 watching

Forks

Releases

Packages

Contributors

Languages