Skip to content

Quickstart

Prerequisites

  • Linux (x86_64, aarch64) or macOS (Apple Silicon / arm64).

1. Install Sequre

/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/0xTCG/sequre/main/scripts/install.sh)"

This installs Codon and Sequre to ~/.sequre and adds it to your PATH.

Manual install 
mkdir -p $HOME/.sequre
curl -L https://github.com/exaloop/codon/releases/download/v0.19.6/codon-$(uname -s | awk '{print tolower($0)}')-$(uname -m).tar.gz \
  | tar zxvf - -C $HOME/.sequre --strip-components=1
curl -L https://github.com/0xTCG/sequre/releases/latest/download/sequre-$(uname -s | awk '{print tolower($0)}')-$(uname -m).tar.gz \
  | tar zxvf - -C $HOME/.sequre
export PATH=$HOME/.sequre/bin:$PATH

2. Run first example

sequre examples/addmul.codon --local

Compilation takes a moment

Sequre programs compile to native machine code, so the first run may take a few minutes. The launcher shows compilation progress by default.

This forks three processes (a trusted dealer + two compute parties) and runs a simple addition-and-multiplication benchmark on secret-shared data.

Expected output: 

CP0:    addmul: 0
CP0:    innerprod: 0
CP1:    addmul: 471
CP2:    addmul: 471
CP1:    innerprod: 32
CP2:    innerprod: 32

The result 7*13 + 13*19 + 7*19 = 471 was computed entirely on secret-shared data — no party ever saw the raw inputs of another.

What just happened?

Here is examples/addmul.codon:

from sequre import main, sequre, Sharetensor as Stensor

@sequre
def addmul(mpc, a, b, c):
    addmul_result = a * b + b * c + a * c
    print(f"CP{mpc.pid}:\taddmul: {addmul_result.reveal(mpc)}")

@sequre
def innerprod(mpc, a, b):
    innerprod_result = a.dot(mpc, b)
    print(f"CP{mpc.pid}:\tinnerprod: {innerprod_result.reveal(mpc)}")

@main
def run(mpc, a, b, c, x, y):
    a_enc = Stensor.enc(mpc, a)
    b_enc = Stensor.enc(mpc, b)
    c_enc = Stensor.enc(mpc, c)
    x_enc = Stensor.enc(mpc, x)
    y_enc = Stensor.enc(mpc, y)

    addmul(mpc, a_enc, b_enc, c_enc)
    innerprod(mpc, x_enc, y_enc)

if __name__ == "__main__":
    # No need to pass mpc argument when calling sequre method with a main decorator
    run(7, 13, 19, [1, 2, 3], [4, 5, 6])

Key concepts:

  1. @main — The entry-point decorator. Pass --local to fork all parties on one machine, or omit it to run in distributed (online) mode.
  2. Stensor.enc(mpc, value) — Secret-shares a plaintext integer into additive shares distributed across parties.
  3. a_enc * b_enc — The @sequre compiler plugin automatically rewrites arithmetic on secret-shared data into Beaver-triple secure multiplications.
  4. .reveal(mpc) — Reconstructs the secret by combining shares from all parties.

Execution modes

@main lets the user control the mode via CLI (--local for local, online otherwise). Sequre also provides lower-level @local and @online decorators for hard-coding the execution mode. See Running Distributed for details.

Run vs. build

The launcher supports both Codon execution modes:

# JIT compile and run immediately (default)
sequre examples/addmul.codon --local

# Compile to a standalone binary
sequre build examples/addmul.codon -o addmul
./addmul --local

Release vs debug mode

Important: Sequre compiles in debug mode by default (with backtraces). Always use -release for production and benchmarks — it is significantly faster.

# Debug mode (default) — slow, with full backtraces on failure
sequre run examples/addmul.codon --local

# Release mode — fast, production-ready
sequre run -release examples/addmul.codon --local

# Building a release binary
sequre build -release examples/addmul.codon -o addmul

Next steps