Introduction to Braid

Braid is a high-performance, statically typed systems programming language that combines Python-like syntax, Rust-like safety, and C-like speed. It targets ML/AI workloads, systems programming, and full-stack web development.

History

Braid was created to bridge the gap between expressive high-level languages and performant low-level systems. The language evolved from a need for a unified toolchain that could handle everything from neural network training to web application development without sacrificing readability or performance. The prototype compiler (braidc) uses a C lexer/parser pipeline with a C++ IR and codegen backend.

Design Philosophy

Braid is built on three pillars:

Python-like syntax — Readable, familiar syntax with braces {} and semicolons ;
Rust-like safety — Strong static typing with Hindley-Milner type inference, ARC with cycle detection
C-like speed — Compilation to native code via C codegen and GCC, with bytecode VM execution

Key Features

Strong static typing — Every value has a type known at compile time
ARC memory management — No garbage collection pauses
First-class tensors — N-dimensional arrays with slicing and device transfer
Diameter construct — Dialectical programming for complex reasoning
@autograd — Built-in automatic differentiation
C FFI — Call C functions directly with extern fn
Match expressions — Pattern matching on integers, strings, booleans, enums, and nil
Bytecode VM — Compile to .bx bytecode for portable execution

Hello World

fn main() {
    print("Hello, Braid!")
}

Variables and Functions

let x = 42
let name = "Braid"
let pi = 3.14

fn add(a: int, b: int) -&gt; int {
    return a + b
}

let result = add(x, 8)
print(result)  // 50

Tensors and ML

let matrix = [[1, 2], [3, 4]]
let val = matrix[0][1]       // 2
let slice = matrix[0:2]      // rows 0-1

@autograd fn loss(x: float, y: float) -&gt; float {
    return (x - y) * (x - y)
}

Diameter Example

diameter temperature: {
    pole hot: {
        return 80.0
    }
    pole cold: {
        return 10.0
    }
}

temperature.evolve()
let tension = temperature.observe("tension")

Installation