ta-tools

A high-performance Technical Analysis library written in Rust and compiled to WebAssembly.

ta-tools provides near-native calculation speeds for technical indicators while remaining universally compatible across Node.js, Bun, Deno, and modern browsers.

Why?

Most existing TA libraries fall into two categories:

1. Pure JavaScript: Easy to use but suffers from garbage collection overhead during heavy calculations.
2. Native C++ Add-ons (e.g., tulind): Fast, but require complex compile-chains (node-gyp) on installation and do not run in the browser.

ta-tools solves this by using WebAssembly. It offers the performance of C++ with the portability of JavaScript.

• Zero Compilation: Users do not need a C++ compiler installed.
• Universal: The exact same package runs on the server and the client.
• Type Safe: Built with Rust for memory safety and correctness.

Installation

npm install ta-tools

Supported Indicators

Moving Averages:

• SMA (Simple Moving Average)
• EMA (Exponential Moving Average)
• WMA (Weighted Moving Average)
• HMA (Hull Moving Average)

Oscillators & Momentum:

• RSI (Relative Strength Index)
• MACD (Moving Average Convergence Divergence)
• Stochastic Fast/Slow
• Stochastic RSI

Volatility:

• Bollinger Bands
• ATR (Average True Range)
• Linear Regression

Trend & Volume:

• ADX (Average Directional Index)
• Ichimoku Cloud
• CVD (Cumulative Volume Delta)
• MFI (Money Flow Index)

Volume Profile & VWAP:

• Fixed Range Volume Profile
• Session VWAP
• Rolling VWAP
• Anchored VWAP

Support Levels:

• Pivot Points (Standard, Fibonacci, Woodie variants)

Usage/API

Basic Usage - Batch Mode

All indicators accept either number[] or Float64Array and return results immediately:

import { sma, ema, rsi, bbands } from 'ta-tools';

const prices = [44.34, 44.09, 44.15, 43.61, 44.33, 44.83, 45.10];

// Simple Moving Average
const sma14 = sma(prices, 14);
console.log(sma14); // Float64Array

// Exponential Moving Average
const ema12 = ema(prices, 12);

// Relative Strength Index
const rsi14 = rsi(prices, 14);

// Bollinger Bands (period, standard deviations)
const bands = bbands(prices, 20, 2);
console.log(bands.upper);    // Upper band
console.log(bands.middle);   // Middle band (SMA)
console.log(bands.lower);    // Lower band
console.log(bands.percentB); // %B indicator

OHLCV Indicators - Candle Input

OHLCV indicators accept Candle[] directly, eliminating positional arguments:

import { atr, adx, ichimoku, mfi } from 'ta-tools';

interface Candle {
  open: number;
  high: number;
  low: number;
  close: number;
  volume?: number;
  time?: number;
}

const candles: Candle[] = [
  { open: 100, high: 102, low: 99, close: 101, volume: 1000000 },
  { open: 101, high: 103, low: 100, close: 102, volume: 1100000 },
  // ... more candles
];

// Average True Range
const atrValues = atr(candles, 14);

// Average Directional Index
const adxResult = adx(candles, 14);
console.log(adxResult.adx);     // ADX line
console.log(adxResult.plusDI);  // +DI
console.log(adxResult.minusDI); // -DI

// Ichimoku Cloud
const cloud = ichimoku(candles, 9, 26, 52);
console.log(cloud.tenkan);   // Conversion line
console.log(cloud.kijun);    // Base line
console.log(cloud.senkouA);  // Leading Span A
console.log(cloud.senkouB);  // Leading Span B
console.log(cloud.chikou);   // Lagging Span

// Money Flow Index
const mfiValues = mfi(candles, 14);

Streaming Mode

For live data, use .stream() to maintain indicator state:

import { rsi, ema, macd } from 'ta-tools';

// RSI streaming
const rsiStream = rsi.stream(14);
rsiStream.init([44.34, 44.09, 44.15, 43.61]); // Initialize with historical data
const currentRsi = rsiStream.next(44.33);     // Add new candle

// EMA streaming
const emaStream = ema.stream(12);
emaStream.init(historicalPrices);
const currentEma = emaStream.next(newPrice);

// MACD streaming
const macdStream = macd.stream(12, 26, 9);
macdStream.init(historicalPrices);
const { macd, signal, histogram } = macdStream.next(newPrice);

Multi-Indicator Analysis

Use analyze() to run multiple indicators on the same data:

import { analyze, sma, ema, rsi, bbands, macd } from 'ta-tools';

const results = analyze(prices, {
  sma20: (d) => sma(d, 20),
  sma50: (d) => sma(d, 50),
  rsi14: (d) => rsi(d, 14),
  bbands: (d) => bbands(d, 20, 2),
  macd: (d) => macd(d, 12, 26, 9),
});

console.log(results.sma20);       // Float64Array
console.log(results.rsi14);       // Float64Array
console.log(results.bbands.upper); // Float64Array
console.log(results.macd.histogram); // Float64Array

Type Utilities

import { toFloat64Array, extractOHLCV } from 'ta-tools';

// Convert number[] to Float64Array
const arr = toFloat64Array([1, 2, 3, 4, 5]); // Returns Float64Array

// Extract OHLCV components from Candle[]
const { open, high, low, close, volume, time } = extractOHLCV(candles);

Performance Notes

ta-tools is optimized to minimize boundary overhead between JavaScript and WebAssembly:

• Zero-Copy Operations: Arrays are passed directly to WASM without copying
• Memory: Uses flat Float64Array buffers to avoid object overhead
• Precision: All calculations use 64-bit floating-point precision (f64)
• Compilation: WASM is pre-compiled and optimized with wasm-opt -Oz

Benchmark Results

Below is a concise, exact summary of the benchmarking results. Each cell shows how ta-tools (WASM) performs relative to the other libraries for the indicated dataset size (values are "times faster" where >1 means ta-tools is faster than the compared library, and <1 means ta-tools is slower).

| Indicator | Small (1k) | Big (10k) | Huge (100k) | Streaming (single next) | |---|---:|---:|---:|---:| | SMA | vs indicatorts: 0.58×vs fast-technical-indicators: 1.82×vs trading-signals: 8.51× | vs indicatorts: 0.69×vs fast-technical-indicators: 2.63×vs trading-signals: 14.14× | vs indicatorts: 1.46×vs fast-technical-indicators: 4.95×vs trading-signals: 13.41× | vs fast-technical-indicators: 0.99×vs trading-signals: 1.16× | | EMA | vs indicatorts: 1.05×vs fast-technical-indicators: 1.27×vs trading-signals: 1.55× | vs indicatorts: 1.29×vs fast-technical-indicators: 1.33×vs trading-signals: 6.08× | vs indicatorts: 2.26×vs fast-technical-indicators: 4.11×vs trading-signals: 6.28× | vs fast-technical-indicators: 0.94× | | RSI | vs indicatorts: 2.52×vs fast-technical-indicators: 16.79×vs trading-signals: 19.71× | vs indicatorts: 2.81×vs fast-technical-indicators: 16.47×vs trading-signals: 19.69× | vs indicatorts: 4.10×vs fast-technical-indicators: 15.71×vs trading-signals: 15.36× | vs fast-technical-indicators: 1.18×vs trading-signals: 1.18× | | WMA | vs fast-technical-indicators: 2.90×vs trading-signals: 9.45× | vs fast-technical-indicators: 4.05×vs trading-signals: 15.11× | vs fast-technical-indicators: 6.48×vs trading-signals: 14.67× | vs fast-technical-indicators: 0.98× | | MACD | vs indicatorts: 0.99×vs fast-technical-indicators: 2.26×vs trading-signals: 4.06× | vs indicatorts: 1.23×vs fast-technical-indicators: 2.65×vs trading-signals: 5.79× | vs indicatorts: 1.45×vs fast-technical-indicators: 4.10×vs trading-signals: 3.64× | vs fast-technical-indicators: 0.81× | | Bollinger Bands | vs indicatorts: 1.08×vs fast-technical-indicators: 6.91×vs trading-signals: 11.52× | vs indicatorts: 1.33×vs fast-technical-indicators: 10.09×vs trading-signals: 16.47× | vs indicatorts: 1.66×vs fast-technical-indicators: 10.69×vs trading-signals: 10.80× | vs fast-technical-indicators: 1.05×vs trading-signals: 1.03× | | ATR | vs indicatorts: 7.13×vs fast-technical-indicators: 1.34×vs trading-signals: 4.42× | vs indicatorts: 9.93×vs fast-technical-indicators: 1.36×vs trading-signals: 5.90× | vs indicatorts: 13.22×vs fast-technical-indicators: 2.85×vs trading-signals: 4.56× | vs fast-technical-indicators: 1.00×vs trading-signals: 1.04× | | Stochastic Oscillator | vs indicatorts: 2.25×vs fast-technical-indicators: 4.80×vs trading-signals: 9.45× | vs indicatorts: 2.00×vs fast-technical-indicators: 3.99×vs trading-signals: 8.08× | vs indicatorts: 2.23×vs fast-technical-indicators: 4.78×vs trading-signals: 8.26× | vs fast-technical-indicators: 0.95×vs trading-signals: 0.73× | | MFI | vs indicatorts: 4.55×vs fast-technical-indicators: 13.57× | vs indicatorts: 6.93×vs fast-technical-indicators: 20.02× | vs indicatorts: 7.53×vs fast-technical-indicators: 14.15× | vs fast-technical-indicators: 1.12× | | ADX | vs fast-technical-indicators: 2.68×vs trading-signals: 8.63× | vs fast-technical-indicators: 3.06×vs trading-signals: 10.06× | vs fast-technical-indicators: 4.90×vs trading-signals: 7.74× | vs fast-technical-indicators: 0.87×vs trading-signals: 0.85× | | StochRSI | vs fast-technical-indicators: 3.25× | vs fast-technical-indicators: 3.54× | vs fast-technical-indicators: 4.30× | vs fast-technical-indicators: 1.10× | | Ichimoku Cloud | vs fast-technical-indicators: 1.79×vs indicatorts: 16.90× | vs fast-technical-indicators: 2.04×vs indicatorts: 98.41× | vs fast-technical-indicators: 1.80× | vs fast-technical-indicators: 1.19× | | VWAP (Session) | vs indicatorts: 0.83× | vs indicatorts: 1.33× | vs indicatorts: 1.31× | - |

*Notes: the ATR rows in the original report list 'ta-tools (WASM) - ATR > Small dataset' as "1.34x faster than fast-technical-indicators" and "7.13x faster than indicatorts" etc. For readability this table shows ta-tools's reported multiplier against each library directly when available; where the report lists another library being faster than ta-tools the table displays the computed inverse (ta-tools relative to that library). Empty cells mean the report did not include a direct entry for that combination.

For full per-indicator, dataset, and streaming breakdowns see benchmarks/2025-12-09-results.md.

Development

Prerequisites

• Rust (1.56+)
• wasm-pack
• Node.js 22+

Build

# Full build (WASM + TypeScript)
npm run build

# WASM only
npm run build:wasm

# TypeScript only
npm run build:ts

Test

npm test
npm run bench

Project Structure

.
├── crates/ta-core/           # Rust WASM library
│   └── src/indicators/        # Indicator implementations
├── js/                        # TypeScript wrapper
│   └── index.ts              # High-level API
├── dist/                     # Compiled TypeScript
├── pkg/                      # Compiled WASM
└── tests/                    # Integration tests

Contributing

Contributions welcome! Please:

1. Fork the repository
2. Create a feature branch
3. Add tests for new indicators
4. Ensure npm run build && npm test passes
5. Open a pull request

Roadmap

• [ ] Add more exotic indicators
• [ ] Browser-specific optimizations
• [ ] Binary persistence/deserialization
• [ ] WebWorker utilities
• [ ] Deno support

License

MIT