ArithmoAI

ArithmoAI is your Gen-Z friendly, quant-powered, AI-supercharged crypto trading companion. Built as a Telegram bot, it unifies multi-exchange portfolio management, AI chat, quantitative metrics, technical analysis, and automated futures strategies into a single, easy-to-use interface.

---

• Demo : https://www.youtube.com/watch?v=N9Xf6F1b_0I
• Presentation : https://docs.google.com/presentation/d/1kzYSd4Zh7j4MuJOomFN1ouMVbSh42LQTVEyTd2r0vW8/edit?usp=sharing

🚀 Features

• Multi-CEX Management

  • Connect Binance (and other CEX) via API keys
  • View spot & futures balances, open orders, positions
  • Place market buy/sell orders directly from Telegram

• Natural-Language Interface (NLP)

  • Classify messages into intents (portfolio_health, market_sentiment, coin_analysis, should_buy, buy_token, sell_token, replace_coin, portfolio_diversity, tax_implications, etc.)
  • Extract symbols & amounts automatically

• QuantMetrics Integration

  • Fetch Sharpe ratio, volatility, max drawdown, daily return from TokenMetrics /quantmetrics API
  • Fetch long-form AI analyses via /ai-reports

• AI Chat & Summaries

  • TokenMetrics AI Chat (/tmai) for open-ended questions (“next 100× coin?”)
  • OpenAI GPT for summarization & enrichment

• Technical Analysis (TA-Lib)

  • ATR (Average True Range) → dynamic stop-loss & take-profit
  • RSI, MACD, Bollinger Bands, MA (SMA/EMA) for entry/exit signals
  • Entry: MA crossovers, volatility breakouts, divergence checks

• Risk & Money Management

  • Stop-Loss: entry_price – 1.5 × ATR
  • Take-Profit: entry_price + 2.0 × ATR
  • Position Sizing via Kelly Criterion:

    f_star = (win_rate * (avg_win_loss + 1) - 1) / avg_win_loss

• Automated Futures Trading

  • Trend-following, mean-reversion hooks
  • Scheduled checks & market orders

• Portfolio Analysis & Rebalancing

  • Real-time USD valuation via CCXT/Binance
  • Allocation %, portfolio Sharpe & volatility, risk score
  • Identify weakest Sharpe coin and recommend replacements from top QuantMetrics picks

• Market Sentiment Dashboard

  • Pull headlines from TokenMetrics sentiments API
  • Summarize & tag tokens as bullish or bearish

• Tax & Compliance Aid

  • Generate tailored tax-reporting guidance based on portfolio composition

---

🏗️ Architecture

User ↔ Telegram Bot ├─ Intent Classification → OpenAI GPT ├─ QuantMetrics APIs → /quantmetrics, /ai-reports, /tmai ├─ CCXT Library → Binance spot & futures ├─ TA-Lib Library → ATR, RSI, MACD, Bollinger, MA… └─ OpenAI → Summarization & Q&A enhancements

🔧 Installation

1. Clone & enter

   git clone https://github.com/yourname/arithmoai.git
   cd arithmoai

2. Create & activate virtualenv

   python3 -m venv venv
   source venv/bin/activate

3. Install dependencies

   pip install -r requirements.txt

4. Environment variables

   Copy .env.example to .env and set:

   TELEGRAM_BOT_TOKEN=your_telegram_token
   OPENAI_API_KEY=your_openai_key
   TOKENMETRICS_API_KEY=your_tokenmetrics_key
   

5. Run

   python bot.py

⚙️ Configuration & Commands

/start

Connect your Binance account if not already connected.
If already connected, it shows demo queries you can try like:

• portfolio health
• buy ETH 10
• replace weakest coin

/cex

List currently connected exchanges and their connection status.

/auto

Trigger the automated futures strategy using quant + TA indicators.

/positions

Display your current open positions (spot + futures).

💬 Usage Examples

Command	Description
portfolio health	📊 Detailed portfolio analytics & risk assessment
market sentiment	📰 News summary + per-token bullish/bearish outlook
analyze BTC	🔍 Coin-specific Quant & AI analysis
should I buy SOL?	✅/❌ Quant signal + metrics breakdown
which coin to buy	✨ Top 5 QuantMetrics picks
buy ETH 10	💸 Market-buy $10 USDT of ETH
sell ADA all	💸 Market-sell your full ADA position
replace weakest coin	🔄 Suggest replacement for lowest Sharpe token
portfolio diversity	🔄 HHI concentration score & diversity rating
tax implications	💰 Custom crypto tax guidance
/auto	🤖 Run automated futures strategy with dynamic leverage
/positions	🤖 Show current futures & spot positions

---

🧠 Technical Deep-Dive

🔌 APIs

TokenMetrics

• GET /quantmetrics?symbol=...
  → Fetches detailed quantitative metrics like Sharpe ratio, volatility, max drawdown, and average daily return.

• GET /ai-reports?symbol=...
  → Returns natural-language investment analysis generated by TokenMetrics AI.

• POST /tmai
  → Sends user prompts to TokenMetrics AI Chatbot for conversational insights and recommendations.

• GET /trading-signals?symbol=...
  → Retrieves AI-generated buy/sell/neutral signals used for auto-futures strategies and short-term trading.

• GET /sentiments
  → Provides aggregated market sentiment and recent news summaries with bullish/bearish classifications.

OpenAI GPT-3.5

• Intent classification from user input
• Summarization of AI insights

CCXT

• Binance spot & futures order placement
• Portfolio balance fetching

---

📊 TA-Lib Indicators

Indicator	Technical Role	Strategy Weightage
ATR	Measures volatility to size stop-loss & take-profit	🟢 High (SL/TP calc)
RSI	Identifies overbought/oversold momentum zones	🟡 Medium
MACD	Confirms trend direction + momentum crossover signals	🟢 High (entry logic)
Bollinger Bands	Volatility envelopes for breakout & mean-reversion	🟡 Medium
SMA/EMA	Defines short/long trend bias and crossover strategies	🟢 High (entry logic)
Stochastic RSI	Combines RSI + momentum for micro timing	🟡 Medium-Low
CCI	Detects price divergence vs mean (momentum filtering)	⚪ Optional
WMA	Weighted moving average for smoother trend reactions	⚪ Optional
ADX	Measures trend strength to filter weak signals	⚪ Optional

⚙️ Only high-confidence indicators (like ATR, MACD, SMA/EMA) are directly used in auto-trading entries and stop logic. Others enhance edge in confluence zones.

📈 Kelly Criterion (Position Sizing)

The Kelly Criterion is a formula used to determine the optimal bet size for maximizing long-term capital growth while minimizing risk. It balances win probability and risk-reward ratio.

Formula:

• f* = (p × R - (1 - p)) / R

Where:

• f* is the optimal fraction of capital to risk per trade
• p is the win probability (success rate)
• R is the risk-reward ratio (average win / average loss)

🔸 How ArithmoAI Applies It

In practice:

• We cap f* at 25% to reduce overexposure

• If the Kelly fraction is negative, we fallback to a safer default stop-loss (e.g., 5%)

• Final stop-loss is calculated as:

• SL = entry_price × (1 - f*) if f* > 0 SL = entry_price × 0.95 if f* ≤ 0

This makes position sizing dynamic and risk-aware, adapting to market volatility and historical performance.

📉 Risk Management

• Entry Signal:

  • Moving average crossover
  • Volatility breakout (Bollinger Band spike)

• Stop-Loss:

  • SL = Entry Price - (1.5 × ATR)

• Take-Profit:

  • TP = Entry Price + (2.0 × ATR)

• Position Sizing (Kelly Criterion):

win_rate     = wins / total_trades
avg_win_loss = avg_profit / avg_loss
f_star       = (win_rate * (avg_win_loss + 1) - 1) / avg_win_loss

---

🔄 Portfolio Rebalancing

• 📥 Fetch current holdings using CCXT
• 💵 Compute USD value and % allocation of each token
• 📉 Rank tokens by Sharpe Ratio to identify the weakest asset
• 🔁 Recommend replacement from the top QuantMetrics picks

---

🤝 Contributing

1. Fork this repository
2. Create a feature branch (git checkout -b feature-name)
3. Implement & test your changes
4. Push your branch (git push origin feature-name)
5. Open a Pull Request describing your enhancement

---

---

📄 License

This project is licensed under the MIT License.

© Aditya Chaplot – Feel free to use, fork, and build on it.

---