@stratchai/indicators

37 technical indicators + 12 series variants for systematic trading. Trend, momentum, volatility, volume, oscillators, chart patterns — zero runtime dependencies, pure functions, TypeScript-typed.

import { calcRSI, calcMACD, calcSupertrend, calcBollingerBands } from "@stratchai/indicators";

const rsi = calcRSI(closes, 14);
// number | null

const macd = calcMACD(closes, 12, 26, 9);
// { macd, signal, histogram, bullish, bearish, ... } | null

const st = calcSupertrend(highs, lows, closes, 10, 3);
// { value, bullish, bearish, distance } | null

const bb = calcBollingerBands(closes, 20, 2);
// { middle, upper, lower, std, width, ... } | null

Install

npm install @stratchai/indicators

Works in any Node 16+ environment. No native compilation, no C bindings, no external runtime dependencies.

Stratchai ecosystem

@stratchai/indicators is the foundation that the other two packages build on:

| Package | Purpose | |---|---| | @stratchai/indicators | 37 indicators + 12 series variants (this package) | | @stratchai/strategy-spec | Declarative strategy specs → generated JavaScript | | @stratchai/backtest | Walk-forward audit primitives for OOS validation |

Use @stratchai/indicators alone for any kind of signal/research code, or together with the other two for a full spec-build-audit loop.

Two APIs: scalar vs series

Each indicator comes in two flavors, optimized for different use cases:

Scalar (latest-value)

Returns one result from a trailing prefix. Use in live agents that call once per bar with the latest data.

import { calcRSI } from "@stratchai/indicators";

const rsi = calcRSI(closes, 14);
// number | null  — RSI computed at the END of the input array

Series (one-per-bar)

Returns an array of length closes.length with the indicator computed at every bar. Use in backtests, charts, or anywhere you need per-bar values.

import { calcRSISeries } from "@stratchai/indicators";

const rsi = calcRSISeries(closes, 14);
// (number | null)[]  — same length as closes; null for the first 14 warmup bars

for (let i = 14; i < closes.length; i++) {
  if (rsi[i] !== null && rsi[i] < 30) {
    /* oversold signal at bar i */
  }
}

Series variants ship for the indicators most commonly used in backtests: calcRSISeries, calcSMASeries, calcEMASeries, calcMFISeries, calcAroonSeries, calcADXSeries, calcSupertrendSeries, calcIchimokuSeries, calcKeltnerSeries, calcOBVSeries, calcBollingerBandsSeries, calcMACDSeries.

Available indicators

Trend / momentum

• calcSMA(prices, period) — Simple moving average
• calcEMA(prices, period) — Exponential moving average
• calcMACD(prices, fast, slow, signal) — Moving Average Convergence Divergence
• calcADX(highs, lows, closes, period) — Average Directional Index (Wilder)
• calcSupertrend(highs, lows, closes, period, mult) — Supertrend (ATR-based trailing band)
• calcParabolicSAR(highs, lows, closes, afStep, afMax) — Parabolic SAR (Wilder)
• calcAlligator(prices, jaw, teeth, lips) — Williams Alligator
• calcAwesomeOscillator(highs, lows, fast, slow) — Bill Williams Awesome Oscillator
• calcROC(prices, period) — Rate of Change
• calcHMA(prices, period) — Hull Moving Average
• calcIchimoku(highs, lows, closes, ...) — Ichimoku Cloud
• calcTrendStructure(closes, highs, lows, opts) — HH/HL trend structure analyzer
• calc52WeekHighLow(closes, highs, lows) — Rolling 52-week extremes

Volatility

• calcBollingerBands(prices, period, k) — Bollinger Bands (J. Bollinger)
• calcATR(highs, lows, closes, period) — Average True Range
• calcATRExpansion(highs, lows, closes, period) — ATR + expansion ratio
• calcKeltner(highs, lows, closes, period, mult, atrPeriod) — Keltner Channels
• calcDonchian(highs, lows, period) — Donchian Channels (Turtle Trading)
• calcMassIndex(highs, lows, period, sumPeriod, bulgeLookback) — Mass Index (reversal pattern)

Volume / flow

• calcOBV(closes, volumes, smaPeriod) — On-Balance Volume
• calcMFI(highs, lows, closes, volumes, period) — Money Flow Index
• calcCMF(highs, lows, closes, volumes, period) — Chaikin Money Flow
• calcVWAP(closes, highs, lows, timestamps, params) — Volume-Weighted Average Price

Oscillators

• calcRSI(prices, period) — Relative Strength Index (Wilder)
• calcStochastic(highs, lows, closes, period) — Stochastic %K
• calcAroon(highs, lows, period) — Aroon Up/Down/Oscillator
• calcVolIndex(prices, k) — In-house volatility index

Chart patterns

• calcHammer(opens, highs, lows, closes, opts) — Hammer candle
• calcEngulfing(opens, highs, lows, closes, opts) — Bullish/Bearish Engulfing
• calcMorningStar(opens, highs, lows, closes, opts) — Morning Star reversal
• calcDoubleBottom(closes, highs, lows, opts) — Double Bottom
• calcCupAndHandle(closes, highs, lows, opts) — Cup & Handle (Bulkowski)
• calcFlagPattern(closes, highs, lows, opts, volumes) — Flag breakout
• calcCandlePattern(opens, highs, lows, closes, params) — 2-candle bullish impulse
• calcAscendingTriangle(closes, highs, lows, opts) — Ascending Triangle

Reference levels

• calcPivotPoints(high, low, close) — Floor pivot points
• calcFibonacci(high, low) — Fibonacci retracements

Conventions

Array order. All inputs are arrays in chronological order — oldest at index 0, most recent at the end. Scalar functions read the most recent values (e.g., prices[prices.length - 1] is the current bar).

Return shape. Most indicators return null when there aren't enough bars to compute (e.g., calcRSI(prices, 14) returns null when prices.length < 15). Composite indicators return objects with named fields ({ value, bullish, expansion, ... }) — see TypeScript types or jsdoc.

No I/O. Every function is pure: same inputs → same output. No filesystem, no network, no time dependency. Easy to test, easy to compose.

Why this library?

Most JavaScript indicator libraries are either old and unmaintained, thin C bindings with platform pain, or return raw numbers without the richer context strategies actually need. Useful patterns — Cup & Handle, Flag, Ascending Triangle, Morning Star — are missing from most libraries entirely.

This library was built to fill three specific gaps:

• Boolean conveniences (adx.trending, supertrend.bullish, band.aboveUpper) — most libraries return raw numbers, leaving threshold logic to every consumer
• Pattern detectors following technical-analysis literature (Bulkowski, Wilder, Bollinger) — Cup & Handle, Flag, Ascending Triangle, Morning Star
• No external dependencies — works in any Node.js environment, no native compilation, no C bindings

License

MIT