ZigX-Python

A maturin-like Python binding system implemented in pure Zig.

ZigX makes it easy to create Python extensions using Zig, providing automatic ctypes-based bindings, type stub generation, GIL support, and cross-platform wheel building.

Note: This Project is in Early Active Development, so there may be Breaking changes!

Features

• 🚀 Pure Zig Implementation - No Python build dependencies beyond standard library
• 📦 Bootstrapped Wheels - Pre-compiled binaries bundled for all platforms
• 🔧 Development Mode - Hot-reload friendly develop command
• 📝 Type Stubs - Automatic .pyi file generation for IDE support
• 🔒 GIL-Safe - Automatic GIL release for ctypes calls (just like maturin/pyo3)
• 🌐 Cross-Platform - Supports Linux (x86_64, aarch64), Windows (x86_64), and macOS (x86_64, arm64)
• 🎯 Automatic Export Detection - No configuration needed, exports are detected from Zig source
• ⚡ uv Integration - Works seamlessly with modern Python tooling
• 💪 No Zig Required - Install and use without needing Zig compiler (binaries pre-built)

Installation

From PyPI (Recommended)

pip install zigx

Or with uv (recommended):

uv pip install zigx

No Zig installation required! ZigX wheels include pre-compiled binaries for your platform.

From Source (Requires Zig 0.14.0+)

git clone https://github.com/muhammad-fiaz/zigx-python.git
cd zigx
uv pip install -e .

Documentation

📚 Full documentation is available at muhammad-fiaz.github.io/zigx-python

The documentation includes:

• Getting started guide
• API reference
• Examples and tutorials
• Troubleshooting guide

Create a New Project

zigx new myproject
cd myproject

This creates a minimal project structure:

myproject/
├── pyproject.toml # Project configuration with zigx build backend
├── src/
│ └── lib.zig # Your Zig code with exported functions
└── myproject/
└── __init__.py # Python package (bindings generated on build)

Write Zig Code

// src/lib.zigconststd=@import("std");
/// Add two integerspubexportfnadd(a: i32, b: i32) i32 {
returna+b;
}
/// Multiply two floatspubexportfnmultiply(a: f64, b: f64) f64 {
returna*b;
}
/// Calculate fibonacci numberpubexportfnfibonacci(n: u32) u64 {
if (n<=1) returnn;
vara: u64=0;
varb: u64=1;
vari: u32=2;
while (i<=n) : (i+=1) {
constc=a+b;
a=b;
b=c;
}
returnb;
}

Development Build

# Build and install in development mode
zigx develop

Use in Python

importmyproject# Call your Zig functions - GIL is automatically released!result=myproject.add(1, 2)
print(f"1 + 2 = {result}")
# With type hints in your IDE!product=myproject.multiply(3.14, 2.0)
fib_10=myproject.fibonacci(10)

Release Build

# Build a release wheel
zigx build --release

This creates a wheel in dist/:

dist/myproject-0.1.0-cp314-cp314-win_amd64.whl

Publish to PyPI

# Build and upload
zigx publish

GIL Support

ZigX provides automatic GIL (Global Interpreter Lock) release just like maturin/pyo3. When you call a native function through ctypes, Python automatically releases the GIL for the duration of the call.

This means your Zig code can run in parallel with other Python threads without any extra configuration:

importthreadingimportmyprojectdefcompute():
# GIL is released during this call - other threads can runresult=myproject.heavy_computation(data)
returnresult# Run computations in parallelthreads= [threading.Thread(target=compute) for_inrange(4)]
fortinthreads:
t.start()
fortinthreads:
t.join()

For explicit GIL control in Zig, you can use the zigx helpers:

constzigx=@import("zigx");
pubexportfnheavy_computation(data: [*]f64, len: usize) f64 {
// GIL is already released by ctypes// Do heavy work without blocking Pythonvarsum: f64=0;
for (0..len) |i| {
sum+=@sin(data[i]) *@cos(data[i]);
}
returnsum;
}

Type Mappings

Zig Type	Python Type	ctypes Type
i8	int	c_int8
i16	int	c_int16
i32	int	c_int32
i64	int	c_int64
u8	int	c_uint8
u16	int	c_uint16
u32	int	c_uint32
u64	int	c_uint64
f32	float	c_float
f64	float	c_double
bool	bool	c_bool
void	None	None
[*]u8	bytes	c_char_p
usize	int	c_size_t

Commands

Command	Description
zigx new <name>	Create a new ZigX project
zigx develop	Build and install in development mode
zigx build --release	Build a release wheel
zigx publish	Build and publish to PyPI
zigx --help	Show help information

Requirements

System Requirements

• Operating System: Linux, macOS, or Windows
• Architecture: x86_64, ARM64

Software Requirements

• Zig Compiler: 0.14.0 or later (0.15.0+ recommended)
  • Download from: https://ziglang.org/download/
• Python: 3.8 or later
• Build Tool: uv (recommended) or pip or hatchling or poetry

Optional Dependencies

For development:

• pytest (testing)
• ruff (linting)
• mypy (type checking)
• mkdocs (documentation)
• hatchling (build backend)

For documentation building:

• mkdocs-material
• mkdocstrings[python]

Documentation

Full documentation is available at muhammad-fiaz.github.io/zigx-python

License

Apache License 2.0 - see LICENSE for details.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Links

• 📚 Documentation
• 🐛 Issue Tracker
• 📦 PyPI Package
• 💻 Source Code