CryptoLens 🔮

An AI-powered crypto research platform combining quantitative signal analysis with Claude-generated analyst briefs.

What it does

• Signal Scorecard — computes RSI, MACD, Bollinger Bands, OBV, and a composite score (-1 to +1) for any tracked coin, plus crypto-specific signals: Fear & Greed index and 24h volume change
• AI Research Briefs — Claude reads the signal data and writes a structured analyst brief with signal summary, bull case, bear case, key levels, and bottom line
• Backtesting Engine — simulate 4 trading strategies with full metrics: Sharpe, Sortino, Calmar, max drawdown, win rate, equity curve vs BTC benchmark
• Strategy Comparison — run all 4 strategies on the same coin and see a ranked leaderboard with combined equity curves

Stack

• Backend: FastAPI + SQLAlchemy + PostgreSQL (Supabase)
• Data: CoinGecko API (price/OHLCV) + Alternative.me (Fear & Greed index)
• Signals: pandas, numpy, ta (technical analysis)
• AI: Anthropic Claude API
• Frontend: Streamlit + Plotly

Architecture

cryptolens/
├── app/
│   ├── api/              # FastAPI route handlers
│   │   ├── coins.py      # Coin management + seeding
│   │   ├── signals.py    # Signal computation endpoints
│   │   ├── backtest.py   # Backtesting endpoints
│   │   └── research.py   # AI research brief endpoints
│   ├── core/             # Config + settings
│   ├── db/               # SQLAlchemy models + session
│   ├── models/           # Pydantic request/response schemas
│   └── services/
│       ├── ingestion.py  # CoinGecko data pipeline
│       ├── signals.py    # Quant signal engine
│       ├── backtester.py # Strategy backtesting engine
│       └── research.py   # Claude-powered research layer
├── frontend/
│   ├── app.py            # Streamlit entry point
│   └── views/            # Research, Backtest, Compare pages
├── scripts/              # CLI tools for seeding and running
└── main.py

Setup

Prerequisites: Python 3.11, PostgreSQL (or Supabase free tier)

# 1. Clone
git clone https://github.com/camczc/cryptolens.git
cd cryptolens

# 2. Create virtual environment
python3.11 -m venv venv
source venv/bin/activate

# 3. Install dependencies
pip install -r requirements.txt

# 4. Configure environment
cp env.example .env
# Edit .env — add DATABASE_URL and ANTHROPIC_API_KEY

# 5. Initialize database
python scripts/init_db.py

# 6. Seed price history (CoinGecko free tier, no API key needed)
python scripts/seed_data.py --coin bitcoin ethereum solana binancecoin

# 7. Compute signals
python scripts/run_signals.py --coin bitcoin ethereum solana binancecoin

Running

# Terminal 1 — API server
uvicorn main:app --reload

# Terminal 2 — Streamlit dashboard
python -m streamlit run frontend/app.py

• Dashboard → http://localhost:8501
• API Docs (Swagger) → http://localhost:8000/docs

API Endpoints

Method	Endpoint	Description
GET	/signals/{coin}	Latest signal scorecard
POST	/signals/{coin}/compute	Recompute signals
GET	/analyze/{coin}	Claude research brief
POST	/backtest	Run a strategy backtest
POST	/backtest/compare	Compare all strategies
GET	/coins	List tracked coins
POST	/coins/{coin}/seed	Seed price history

Strategies

Strategy	Logic	Best for
Composite	Weighted aggregate of RSI + MACD + Bollinger + Fear&Greed	General use
RSI Mean Reversion	Buy RSI < 30, sell RSI > 70	High-volatility crypto
Golden Cross	SMA50 vs SMA200 crossover	Trending markets
Fear & Greed Contrarian	Buy extreme fear (< 25), sell extreme greed (> 75)	Sentiment-driven moves

Environment Variables

DATABASE_URL=postgresql://...     # PostgreSQL connection string
ANTHROPIC_API_KEY=sk-ant-...      # Required for AI research briefs
ENV=development
LOG_LEVEL=INFO

Data Sources

• CoinGecko — OHLCV price history, market cap (free, no API key required)
• Alternative.me — Fear & Greed index history (free)

Demo

Research & Signals	AI Research Brief

Backtest Results	Strategy Comparison