Trading Engine

Version 0.4.0 — AI-Augmented Quantitative Trading Platform

Hybrid system combining classical quant signals, LLM agent reasoning, comprehensive risk management, and a feature-rich desktop-grade frontend.

Quick Start

# Install
pip install -e ".[dev,llm,live,ml]"
cd frontend && npm install

# Run (two terminals)
python scripts/dashboard.py          # API on :8000
cd frontend && npm run dev           # UI on :5173

# Or with Docker
docker compose up

Architecture

+------------------------------------------------------------------+
|                   Frontend (React + TypeScript)                    |
|  Dashboard | Portfolio | Signals | Trades | Chart | Backtest     |
|  Agents | Hedge Flow | Strategy Lab | AI Features | Right Sidebar|
+------------------------+-----------------------------------------+
                         | REST + SSE + WebSocket
+------------------------v-----------------------------------------+
|              FastAPI REST API (60+ route modules)                 |
|  Portfolio | Signals | Backtest | Agents | LLM | FinScript       |
|  AI: Briefing | Co-Movement | Earnings | Strategy | Indicator    |
|  Chart Inspector | LLM Query | Streaming SSE                     |
+--+----------+----------+----------+-----------------------------+
   |          |          |          |
+--+--+  +----+----+  +--+---+  +--+--------------+
|Signals|  |Agents |  | Risk |  | Execution   |
|23+   |  |16 HF  |  |Engine|  | Backtest    |
|Engines|  |8 Quant|  |Limits|  | Paper       |
|Regime |  |8 LLM  |  |Stops |  | Alpaca      |
|ML     |  |Renaiss|  |Sizing|  | CCXT / IBKR |
+------+  +-------+  +------+  +-------------+

Key Features

Area	Capabilities
Signals	SMC (order blocks, FVGs, BOS/CHOCH), harmonics, H&S, flags/pennants, price action, regime detection (trend, vol, Wasserstein), ML pattern mining, vision-based detection
Agents	16 hedge fund personas (Buffett, Burry, Graham, Taleb, etc.), 8 quant agents, 8 LLM agents, Renaissance-style teams, debate system
Risk	Position limits, ATR stop-loss, Kelly sizing, circuit breakers, composite risk engine
Backtesting	Event-driven engine, Monte Carlo (1000+), walk-forward, scenario stress tests, synthetic data
FinScript	Custom trading DSL: lexer, parser, AST, interpreter, 40+ built-in functions, PineScript/MT5/TDX export
Analytics	22+ metrics (Sharpe, Sortino, Calmar, VaR, CVaR), attribution, CFA toolkit
Data	yfinance, OpenBB, Alpaca, CCXT, Databento, FRED, Finnhub, SEC, news, Twitter, World Bank
Execution	Backtest, paper trading, Alpaca, CCXT (100+ exchanges), Interactive Brokers
API	FastAPI, 60+ REST endpoints, SSE streaming, WebSocket, JWT auth, API key auth
Frontend	30-page SPA, TradingView charts, React Flow, Zustand, 7 themes, multi-window, distraction-free mode
AI Features	AI Briefing, Ask the Terminal, Strategy Generator, Indicator Generator, Chart Inspector, News Co-Movement, Earnings Call Summary
UI/UX	High-contrast/sunlight themes, distraction-free mode, drag-and-drop (symbol, price, date), multi-window sync, right sidebar (news/calendar/chat)

CLI Commands

# Backtest
trading-engine-backtest --tickers AAPL,MSFT --start 2024-01-01

# Live paper trading
trading-engine-live --mode paper --tickers BTC/USD

# Dashboard API server
trading-engine-dashboard --port 8000

Project Structure

core/           — Types, enums, events, errors
data/           — Market data providers + cache + realtime
signals/        — Signal pipeline (23+ engines, regime, ML, vision)
agents/         — Classical quant, LLM, hedge fund persona, Renaissance
risk/           — Position limits, stop-loss, sizing, circuit breakers
execution/      — Backtest, paper, Alpaca, CCXT, IBKR
backtesting/    — Engine, Monte Carlo, walk-forward, scenario, synthetic
analytics/      — Metrics, attribution, reports, CFA, visualization
api/            — FastAPI REST + SSE + auth
  routes/       — 60+ route modules (including 7 new AI endpoints)
    llm.py            — LLM completion + streaming SSE
    briefing.py       — AI Briefing on demand
    network_co_movement.py — News-driven correlation analysis
    earnings_summary.py   — Earnings call bull/bear/risk summary
    ai_strategy.py        — Natural language → FinScript strategy
    ai_indicator.py       — Natural language → custom indicator
    ai_inspector.py       — Chart pattern LLM analysis (streaming)
    llm_query.py          — Portfolio-aware natural language query
finscript/      — Custom trading DSL (lexer/parser/interpreter)
integrations/   — Discord, Slack, Telegram, SMS, Email, TradingView
persistence/    — SQLAlchemy async ORM, Alembic migrations
frontend/       — React/TypeScript SPA (30 pages)
  src/
    components/
      rightsidebar/  — News, Calendar, Chat, Co-Movement, Earnings tabs
      AIBriefing.tsx       — Modal briefing overlay
      NewsCoMovement.tsx   — News-driven co-movement panel
      EarningsSummary.tsx  — Earnings call summary panel
      StrategyGenerator.tsx — NL → FinScript code gen + backtest
      IndicatorGenerator.tsx — NL → JS indicator plugin gen
      AIInspector.tsx      — Streaming pattern analysis modal
      StreamResponse.tsx   — Reusable SSE streaming text component
    contexts/
      DistractionFreeContext.tsx — Hide chrome for focus mode
      MultiWindowContext.tsx     — BroadcastChannel cross-tab sync
      SymbolDragContext.tsx      — Native HTML5 DnD for symbols
    utils/
      broadcastChannels.ts — Channel name constants + getTabId()
      tickSound.ts         — Web Audio API tick chirp
    api/
      llm.ts               — All LLM + AI API functions
config/         — Settings + strategy defaults

Changelog — 3 June 2026

New AI Features (7 endpoints + 7 components)

AI Briefing — Button-driven market + portfolio briefing with LLM-generated analysis (portfolio, regime, top movers, risk)
Ask the Terminal — LLMPanel now has a "Data Query" mode that injects portfolio/risk/trade context into prompts
AI Strategy Generator — Describe a strategy in plain English → get FinScript code → review → run backtest
AI Indicator Generator — Describe an indicator → get JavaScript code → add to chart at runtime
AI Chart Inspector — Click "What is this?" on a detected pattern → streaming LLM analysis (explanation, analogs, trading implications)
News Co-Movement — Right sidebar tab: enter headline + tickers → AI shows correlated movers
Earnings Call Summary — Paste transcript → AI extracts bull case, bear case, and single biggest risk

New UI/UX Features

Multi-Window — Ctrl+N / Ctrl+Shift+N opens new windows; theme and symbol changes sync via BroadcastChannel
Drag Symbol Anywhere — Drag from SymbolSearch → drop on chart, order entry, compare, or widget
Drag Price to Alert — Drag a price level from the chart → drops on Alert button to pre-fill
Drag Date to Backtest — Drag a date from the chart → drops on backtest start date input
Distraction-Free Mode — Ctrl+Shift+D hides all chrome; floating "Exit Focus" button
High-Contrast / Sunlight Themes — Two new themes: max-contrast black/white and warm sunlight-optimized
Right Sidebar — Collapsible News, Calendar, Chat tabs with keyboard shortcut Ctrl+Shift+R

Streaming Infrastructure

/llm/complete-stream — SSE endpoint for token-by-token LLM responses (used by Chart Inspector)
llmCompleteStream() — Frontend ReadableStream reader

Environment Variables

See .env.example for all config vars. Key ones:

Variable	Required	Default	Purpose
`TRADING_ENGINE_API_KEY`	No	—	API key auth for REST endpoints
`JWT_SECRET_KEY`	No	(auto-generated)	JWT signing secret
`CORS_ORIGINS`	No	`localhost:5173`	Allowed CORS origins
`FINNHUB_API_KEY`	No	—	Market data (quotes, news, search)
`OPENAI_API_KEY`	For AI features	—	OpenAI LLM provider
`ANTHROPIC_API_KEY`	For AI features	—	Anthropic LLM provider

Testing

# Backend
pytest --cov=.

# Frontend
cd frontend && npm run test

Documentation

Guide	Description
User Guide	Complete guide to every page, feature, and button
Quick Start	Installation and first steps
Troubleshooting	Fix common issues step by step
Feature Reference	Every feature explained
Keyboard Shortcuts	All keyboard shortcuts
Architecture	System architecture
API Reference	REST API documentation
Signal Engines	Signal generation docs
AI Agents	Hedge fund personas and LLM agents
Risk Management	Risk engine documentation
Backtesting	Backtesting engine docs
FinScript	Custom trading DSL docs

License

MIT — see LICENSE for details.

ShaikhWarsi/Project_XHASH

README