Crypto AI Trader: Next-Gen Machine Learning Trading & Bot Operations Platform

A complete, self-hosted cryptocurrency trading platform that bridges deep learning price forecasting, real-time Binance execution, mobile Telegram interaction, and multi-client SaaS capability.

???? The Vision

In the highly volatile crypto market, manual trading is emotionally taxing, while typical algorithmic bots lack adaptability. Crypto AI Trader solves this by combining TensorFlow/Keras-based prediction models with a unified mobile interface (Telegram) and a local web operations dashboard (Orchestration Bridge).

Whether you want to run risk-free simulations in a sandbox, manage your own live funds via isolated margin, or orchestrate trading signals for multiple clients, this platform provides a secure, modular, and robust solution.

???? Key Pillars & Features

1. ???? Deep Learning Predictive Models

At the heart of the platform is a machine learning pipeline that processes real-time Binance market data to forecast market trends.

• Multi-Timeframe Feature Engineering: Models analyze technical indicators (RSI, oscillators, price volatility) across historical klines.
• Diverse Model Architectures: Supports multiple neural network versions including Dense chart models and LSTM (Long Short-Term Memory) networks for sequential time-series data.
• Auto-Rebalancing: The AI decision engine runs periodic cycles (evaluating positions every 5 minutes and running full strategy iterations every 30 minutes) to adjust leverage targets based on trend confidence and price momentum.
• Core Implementation: Check out the [Trainer package](about:blank) and the neural network loading module [bot.py](about:blank).

2. ???? Dual-Mode AI Trading Engine

Execute trades with precision and custom risk management in two distinct modes, operating under independent settings:

• Paper Trading (Sandbox): A simulated wallet starting with a virtual balance of 10,000 USDT to benchmark neural network performance, calculate fees, and monitor liquidations with zero risk.
• Live Trading: Real-time trade execution on Binance using isolated margin to maximize capital efficiency while strictly controlling downside risk.
• Safety Protocols: Integrates position watchdog timers, leverage limits (1x to 10x long, 1x to 2x short), daily loss caps, and circuit breakers that halt the bot in the event of persistent API connection drops.
• Core Implementation: Explore the [AIBotService](about:blank) and the Binance integration client [trader.py](about:blank).

3. ???? Telegram Command Center (Legacy)

The Telegram interface is now considered a legacy entry point, having been succeeded by the web-based administrative dashboard which offers significantly faster loading times and more precise data synchronization. While the web dashboard is the primary tool for platform management, Telegram remains fully operational as a convenience for users who wish to check status or execute quick commands without undergoing the standard web login process.

• Interactive Trading Flow: Place market, limit, or OCO (One-Cancels-Other) orders on the fly, with automated limit sells placed in profit-taking ratios upon order execution.
• Trade Signal OCR: Upload screenshots of trading charts or signals from other groups; the bot automatically extracts text, parses parameters, and initiates interactive trade confirmations.
• Real-time Notifications: Instant price alerts, news delisting notifications, and daily PnL summaries delivered straight to your chat.
• Core Implementation: See [photo_handler.py](about:blank) for image parsing and the main entry point [app.py](about:blank).

4. ????️ Orchestration Bridge: Local Operations Dashboard

A self-hosted web console built for managing the bot on multi-cluster orchestration environments with Kubernetes and Docker.

• State Management: View and modify bot configurations (active symbols, leverage limits, notification preferences) and apply updates dynamically.
• Ops Command Center: Run remote services (systemd control), track logs (journalctl and deploy logs), and trigger secure git deployment pulls.
• Production-Grade Security: Bound locally by default and secured via HTTP Basic Auth, with production options for Cloudflare Tunneling, mandatory 2FA, and strict API rate limits.
• Core Implementation: Explore the FastAPI dashboard codebase in [tools/orchestration_bridge/](about:blank).

5. ???? BotTrader: Client Copy-Trading SaaS

Turn your trading signals into a subscription service:

• Subscription Management: Clients subscribe to receive signals, with automated USDT/BTC deposit monitoring extending their subscription periods.
• Secure Client Key Storage: Clients input their API keys directly through Telegram. The bot immediately encrypts these keys at rest using an AES/RSA master key configuration and redacts them from log files.
• Managed Mirroring: Executes bot signals across client sub-accounts simultaneously with individual custom leverage ceilings.
• Core Implementation: Review [routes/bottrader.py](about:blank).

## ????️ The Tech Stack

• Programming Language: Python 3.8+ (type-hinted, modular structure)
• Machine Learning: TensorFlow / Keras (dense & LSTM models)
• Backend Frameworks: FastAPI (operations dashboard), python-telegram-bot (communication pipeline)
• Databases: SQLite (historical market klines and Telegram state), JSON/Pickle (bot settings and wallets)
• Data Sources: Binance WebSockets & REST API
• Security & Ops: systemd, Git hooks deployment, RSA encryption, Cloudflare Tunnels
• Dependency Injection: Unified [AppContainer](about:blank) managing service lifecycles.

???? Security & Safety First

• Log Redaction: All incoming API keys or sensitive data are instantly redacted before writing to log files.
• Permission Hierarchy: Strictly separates command access (Owner/Admin, Authorized Traders, and BotTrader Clients) to prevent unauthorized trades or configurations.
• Local-First Architecture: Keeps your data and secrets on your own hardware, running under your control.