MPLang: A Programming Language for Multi-Party Computation

MPLang is a Python-native library for building and executing multi-party and multi-device programs. It simplifies secure computation by allowing developers to write a single program that orchestrates multiple parties in a synchronous, SPMD (Single Program, Multiple Data) fashion.

Features

Single-Controller SPMD: Write one program that runs across multiple parties in lockstep.
Explicit Device Placement: Clearly annotate and control where data lives and computation happens (e.g., on party P0, P1, or a secure SPU).
Function-Level Compilation: Use the @mplang.function decorator to compile Python functions into an auditable, optimizable graph representation.
Pluggable Architecture: Easily extend MPLang with new frontends (like JAX, Ibis) and backends (like StableHLO, SPU).

Getting Started

Installation

You'll need a modern Python environment (3.10+). We recommend using uv for fast installation.

# Install uv (if not already installed)
curl -LsSf https://astral.sh/uv/install.sh | sh

# Install MPLang from PyPI
uv pip install mplang

Quick Example

Here's a taste of what MPLang looks like. This example shows a "millionaire's problem" where two parties compare their wealth without revealing it.

import mplang
import mplang.device as mpd
from numpy.random import randint

# Use a decorator to compile this function for multi-party execution
@mplang.function
def millionaire():
    # Alice's value, placed on device P0
    x = mpd.device("P0")(randint)(0, 1000000)
    # Bob's value, placed on device P1
    y = mpd.device("P1")(randint)(0, 1000000)
    # The comparison happens on a secure device (SPU)
    z = mpd.device("SPU")(lambda a, b: a < b)(x, y)
    return z

# Set up a local simulator with 2 parties
sim = mplang.Simulator(2)

# Evaluate the compiled function
result = mplang.eval(sim, millionaire)

# Securely fetch the result
print("Is Alice poorer than Bob?", mplang.fetch(sim, result))

Learn More

Tutorials: Check out the tutorials/ directory for in-depth, runnable examples covering conditions, loops, and more.
Contributing: We welcome contributions! See our Contributing Guide to get started with the development setup.

License

MPLang is licensed under the Apache 2.0 License.

Metadata

Release files for mplang-nightly 0.1.dev143

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for mplang-nightly 0.1.dev143
File	Size	Uploaded
mplang_nightly-0.1.dev143.tar.gz	307.4 kB	Sep 26, 2025	Details

Built distribution (wheel)

Table of built distributions (wheels) for mplang-nightly 0.1.dev143
File	Interpreter	ABI	Platform	Reset
mplang_nightly-0.1.dev143-py3-none-any.whl 200.7 kB Sep 26, 2025	Python 3	none	any	Details

Total release size: 508.1 kB

Release files / mplang_nightly-0.1.dev143.tar.gz

Download URL	mplang_nightly-0.1.dev143.tar.gz
Size	307.4 kB
Tags	Source
SHA-256 checksum How to use checksums	`0e67e7ab87714581c68d1f97cfab463736dba1a9ee4df96c39d46a043b024b1f`
BLAKE2b-256 checksum How to use checksums	`a408083df82cd12b603c19fbaa24e89dff8cc05c9efb1ed225543362d616a128`
Upload date	Sep 26, 2025
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`python-httpx/0.28.1`

Release files / mplang_nightly-0.1.dev143-py3-none-any.whl

Download URL	mplang_nightly-0.1.dev143-py3-none-any.whl
Size	200.7 kB
Tags	Python 3
SHA-256 checksum How to use checksums	`7a32f19abcf43a4847b14bb7089ac3bd32fd5a758c4fc90250b21a68d9af1658`
BLAKE2b-256 checksum How to use checksums	`fad44412f35723de0c0f87c0811170e10f8478640d1c14fb59ec50ea94273fb7`
Upload date	Sep 26, 2025
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`python-httpx/0.28.1`