Splitr

Splitr is a minimalist TypeScript library designed for highly accurate traffic splitting and outcome distribution. With near 100% precision and zero dependencies, it provides a reliable solution for A/B testing, feature flagging, or any scenario requiring weighted random selection.

Key Features

• High Precision: Achieves 99.95% - 99.97% accuracy in split distributions
• Zero Dependencies: Completely standalone with no external requirements
• TypeScript Native: Built with full type safety and modern TypeScript features
• Minimalist API: Simple, intuitive interface for defining and executing splits
• Flexible Usage: Perfect for A/B testing, traffic distribution, or any weighted selection needs

How It Works

Splitr uses a weighted distribution algorithm to ensure outcomes match their specified probabilities with extremely high accuracy. Each option is assigned a weight (percentage), and the sum of all weights must equal 100.

Example Usage

Here’s a simple example demonstrating traffic splitting:

const options = [
    {
        weight: 10,
        value: "10% of the time",
    },
    {
        weight: 20,
        value: "20% of the time",
    },
    {
        weight: 70,
        value: "70% of the time",
    },
];

const splitr = new Splitr(options);
const result = splitr.run();
// Returns one of the values based on their weights

Splitr’s combination of simplicity, accuracy, and zero dependencies makes it an excellent choice for developers who need reliable outcome splitting without the overhead of larger libraries.

GitHub