Production-grade algorithmic trading infrastructure for Node.js — from Pine Script execution and LLM signal generation to live exchange connectivity and interactive dashboards. One codebase for backtest, paper, and live trading.

Start here: Clone the reference implementation — a fully working news-sentiment AI trading system with LLM forecasting, multi-timeframe data, and a documented February 2026 backtest.

# Scaffold a new AI agent trading project
npx -y @backtest-kit/cli --init --output my-trading-bot
cd my-trading-bot
npm start

Look-Ahead Bias is Architecturally Impossible

Temporal context flows automatically through AsyncLocalStorage — every getCandles() call is implicitly bounded to the current backtest tick. No timestamp parameters to pass, no future data to accidentally leak.

Same Code, Backtest → Paper → Live

Strategy functions are unaware of execution mode. Switch from historical simulation to real exchange connectivity by changing one flag.

LLM as Signal Generator

The framework provides structured pipelines for injecting multi-timeframe technical analysis, order book data, and news sentiment into LLM context — while keeping temporal isolation and validation invariants intact.

Pine Script indicators if you want to

Port your existing TradingView strategies to Node.js without rewriting a single line — @backtest-kit/pinets executes Pine Script v5/v6 locally via the PineTS runtime, extracts plot values into typed objects, and feeds them directly into the signal pipeline alongside LLM-generated signals.

A series of in-depth articles documenting the engineering decisions behind the framework:

1. Look-Ahead Bias — How AsyncLocalStorage makes temporal contamination architecturally impossible
2. Second-Order Chaos — Why thousands of identical bots trade against themselves at a loss
3. Claude Trader — How AI gets hands for autonomous strategy iteration via Claude Code
4. Option Hedging — Why the price drops in a single candle and how to adapt
5. AI Strategy Workflow — AI workflow for identifying and updating liquidation cascade criteria
6. AI Strategy Blueprint — ReAct-pattern LLM trading agent with live test results
7. AI News Trading Signals — News sentiment analysis as a trading signal: methodology and backtest
8. AI Liquidity Harvesting — How to detect and invert manipulative Telegram signal channels
9. Pine Script Local Markets — Running Pine Script on exchanges not available in TradingView
10. DCA Averaging Strategy — Asset price averaging mechanics, drawdown reality, and broker deception

Architectural and infrastructure concepts behind the framework:

1. Monorepo Parallel Execution — Declarative event-driven strategies in a monorepo, N entry points per mode, ~700× replay speed per symbol
2. Zero Expectation Escape — Why every strategy decays to zero EV and how to trade on external capital inflow instead
3. Manual Control Adapter — Debugging the exchange adapter in an hour via dashboard buttons that pull the same hooks as the strategy
4. Highload Candle Storage — MinIO vs MongoDB vs PostgreSQL + Pgpool-II for 260M candles, and what Redis lookup actually buys
5. AI Signal Scoring Agent — MCP agent that trades Telegram signals in paper mode and rates authors by real Sharpe/Calmar, not their screenshots

• 📖 Documentation
• 💻 GitHub Repository
• 🎯 Reference Implementation
• 🤖 DeepWiki

_{MIT © tripolskypetr}