Open-source trading intelligence engine for humans and AI agents.
Pass OHLC bars in → get market structure and a scored setup JSON out.
This is a local npm library. It is not a hosted API, not a broker, and not “another RSI package.”
Commodity indicators (RSI, MACD, SMA) are everywhere. What agents actually need is deterministic structure:
The LLM explains facts — it does not invent whether an FVG exists.
npm install @vaultcharts/trading-engine
Requires Node.js 18+. Runs on your machine / in an agent process. No VaultCharts server.
import { analyzeSetup } from '@vaultcharts/trading-engine';
const result = analyzeSetup({
symbol: 'BTCUSDT',
timeframe: '4h',
data: ohlcBars, // { time, open, high, low, close, volume? }[]
});
console.log(result);
// {
// trend: 'bullish',
// bos: { direction: 'bullish', level: ... },
// fvg: { bullish: true, mid: ... },
// liquiditySweep: { ... },
// score: 84,
// entry, stop, targets, riskReward,
// evidence: ['...']
// }
Optional. The same package can power a cloud endpoint later for people who don’t want to run Node. Phase 1–2 do not require a server.
cd trading-engine
npm install
npm test
npm run demo
Vitest covers FVG detection, risk/reward math, and analyzeSetup / analyzeStructure on synthetic OHLC (deterministic scoring). This is a starter suite — not full coverage of every primitive yet.
MIT
| In (Phase 1) | Later |
|---|---|
Structure primitives + analyzeSetup | Full pattern-detector port |
| Local npm / agent use | scanUniverse watchlist filter |
| MIT | Desktop remains reference client |
| S&D MTF stubbed empty | Wire real multi-TF S&D |
Website (vaultcharts.com) stays: download desktop + free tools + docs pointing at this engine.
2 commits
TypeScript
100.0%
Open-source trading intelligence engine for humans and AI agents.
Pass OHLC bars in → get market structure and a scored setup JSON out.
This is a local npm library. It is not a hosted API, not a broker, and not “another RSI package.”
Commodity indicators (RSI, MACD, SMA) are everywhere. What agents actually need is deterministic structure:
The LLM explains facts — it does not invent whether an FVG exists.
npm install @vaultcharts/trading-engine
Requires Node.js 18+. Runs on your machine / in an agent process. No VaultCharts server.
import { analyzeSetup } from '@vaultcharts/trading-engine';
const result = analyzeSetup({
symbol: 'BTCUSDT',
timeframe: '4h',
data: ohlcBars, // { time, open, high, low, close, volume? }[]
});
console.log(result);
// {
// trend: 'bullish',
// bos: { direction: 'bullish', level: ... },
// fvg: { bullish: true, mid: ... },
// liquiditySweep: { ... },
// score: 84,
// entry, stop, targets, riskReward,
// evidence: ['...']
// }
Optional. The same package can power a cloud endpoint later for people who don’t want to run Node. Phase 1–2 do not require a server.
cd trading-engine
npm install
npm test
npm run demo
Vitest covers FVG detection, risk/reward math, and analyzeSetup / analyzeStructure on synthetic OHLC (deterministic scoring). This is a starter suite — not full coverage of every primitive yet.
MIT
| In (Phase 1) | Later |
|---|---|
Structure primitives + analyzeSetup | Full pattern-detector port |
| Local npm / agent use | scanUniverse watchlist filter |
| MIT | Desktop remains reference client |
| S&D MTF stubbed empty | Wire real multi-TF S&D |
Website (vaultcharts.com) stays: download desktop + free tools + docs pointing at this engine.
2 commits
TypeScript
100.0%