A daily pre-market screening and recommendation system. Given a US-equity universe, it computes a multi-signal Heat Score, identifies Smart Money flow, scores Sentiment, applies a multi-signal resonance recommendation engine, and ships the result as a Telegram report, Markdown archive, JSON watchlists, and a web dashboard.
Companion to AI_trader. AI_trader executes; Market Radar finds targets. The two communicate via
data/proposed_*.jsonwatchlists.
| Pillar | Signals |
|---|---|
| 🔥 Heat Score | volume × ADV, large-block %, UOA count, 24h news density |
| 💰 Smart Money | block trade buy/sell imbalance, UOA call vs put, P/C ratio, IV skew (25Δ) |
| 📈 Technicals | RSI(14), MACD(12/26/9), SMA 20/50/200, Bollinger Bands, ATR(14), 52w high/low |
| 📊 Sentiment | Claude Haiku 4.5 news headline scoring (with prompt caching), StockTwits Bullish/Bearish |
| 🏛️ Insider | SEC EDGAR Form 4 — open-market buys vs sells, weighted (P codes 2× weight) |
| 🎯 Recommendations | multi-signal resonance + risk veto (strong_long / watch / avoid) |
| 📤 Outputs | Telegram + reports/YYYY-MM-DD.md + watchlists for AI_trader + web dashboard |
All 9 phases of the original plan are complete. See TARGET.md for the full status; see USER_GUIDE.md for operational instructions (中文版).
| Phase | Status |
|---|---|
| 0 — Project init | ✅ |
| 1.1–1.4 — Data pipelines (DB, screener, news, blocks) | ✅ |
| 2 — Technicals | ✅ |
| 3 — Heat + first report ⭐ MVP | ✅ |
| 4 — Options Flow ⭐⭐ | ✅ |
| 5 — Sentiment ⭐⭐⭐ | ✅ |
| 6 — Engine v2 + backtest | ✅ |
| 7 — Automation (LaunchAgent + Telegram) | ✅ |
| 8 — SEC EDGAR insider trades | ✅ |
| 9 — Web Dashboard | ✅ |
# 1. Create venv (Python 3.10+ required)
python3.11 -m venv .venv
source .venv/bin/activate
# 2. Install
pip install -e ".[dev,dashboard]"
# 3. Configure
cp .env.example .env
# edit .env — add ALPACA, TELEGRAM, ANTHROPIC keys
# 4. Initialize the database
radar db init
# 5. Sanity check
radar ping
# → Market Radar OK
# 6. Run the daily pipeline manually
bash scripts/run_daily.sh
# 7. Start the dashboard
radar dashboard --port 8765
# open http://localhost:8765 Alpaca SEC EDGAR StockTwits Anthropic
│ │ │ │
▼ ▼ ▼ ▼
screener · news · trades · option chain · Form 4 · sentiment LLM
│
▼
SQLite (data/radar.db, 13 tables)
│
▼
Heat / Smart Money / Sentiment / Insider scoring
│
▼
Engine v2 — multi-signal resonance + risk veto
│
├──▶ Telegram (daily + 6× hourly intraday)
├──▶ reports/YYYY-MM-DD.md
├──▶ data/proposed_*.json (→ AI_trader)
└──▶ FastAPI Dashboard (http://localhost:8765)
Python 3.10+ · alpaca-py · ta (technical-indicator library) · py_vollib · SQLite · loguru · pydantic-settings · click · Anthropic SDK (Haiku 4.5 with prompt caching) · FastAPI + uvicorn · macOS launchd · Telegram Bot.
- USER_GUIDE.md — full English operations guide
- USER_GUIDE.zh-TW.md — 繁體中文操作指南
- DISCLAIMER.md — financial / trading disclaimer (bilingual). Read this before using.
- TARGET.md — phase-by-phase implementation history
- SAVE.md — session journal (gitignored, local only)
Licensed under the PolyForm Noncommercial License 1.0.0.
In short: free for personal, research, educational, and noncommercial use. Commercial use is not permitted without a separate license. See LICENSE for the binding terms.
This is a research and educational tool. It is not investment advice; the author is not a registered investment advisor; past or backtested performance does not predict future results; you assume all trading risk. The full disclaimer is in DISCLAIMER.md. By using this software you accept those terms.