Jac and Jaseci

The Jac programming language and Jaseci runtime stack features Python-like syntax, compiles to Python bytecode, JavaScript, and native machine code, and makes meaning types, Object-Spatial Programming, and scale invariance language features.

Get Started Try Playground

Our journey is strengthened by our sponsors and partners

Part of Nvidia Inception Program, supporting cutting edge AI innovation and scalability

Research led by faculty and researchers at University of Michigan, United States

Sponsored by NSF, advancing community driven open source ecosystem

What is Jac?

Jac features Python-like syntax and compiles to Python bytecode, JavaScript, and native machine code (C-ABI compatible). Every library in PyPI, npm, and native C ecosystems works directly -- no interop wrappers needed. Jac introduces new features to minimize complexity and accelerate AI application development. We also provide library mode.

awesome.jac

Output:

Get Started with Jac

🚀

Getting Started

Install Jac and setup your development environment. Write your first Jac program.

Start Guide

🎯

Tour of Jac

Learn core concepts including Object-Spatial Programming and AI constructs.

Take Tour

📚

Language Reference

Complete guide for Python developers transitioning to Jac programming.

Read Reference

🐦

AI Day Planner

Build a complete full-stack AI app in one guided project. The flagship tutorial.

Build the App

🤖

Agentic AI

Give an LLM tools it can call: build an agent that reads and acts on your data.

Build an Agent

🎮

Build a Chess Engine

The native pathway hands-on: from Jac source to a standalone compiled binary.

Go Native

☁️

Jac Scale

Deploy your Jac applications to the cloud with built-in scaling and persistence.

Go Scale

🧠

AI Programming

Master AI integration with byLLM and multimodal models in Jac.

AI Guide

⚡

Quick Reference

Jac in 2 minutes! Quick syntax reference and cheat sheets for fast learning.

Quick Start

Why Choose Jac?

🐍

Ecosystem-Native Compilation

With Python-like syntax, Jac compiles to Python bytecode, JavaScript, and native machine code -- every library just works.

🧠

Meaning-Typed AI Integration

Jac introduces a new way to integrate LLMs into code: replace a function body with by llm(), and the prompt is derived from your names, types, and sem annotations. There is no prompt to write.

🌐

Object-Spatial Programming

Nodes, edges, and walkers make the moving locus of computation a language construct: computation travels through your data's topology, dispatched by arrival.

☁️

Scale Invariance

Jac makes persistence and user concepts part of the language, so the same program text runs unchanged in a terminal or across a cluster.

⚡

One-Command Deployment

Deploy without writing infrastructure. jac start --scale provisions Kubernetes, Redis, and MongoDB, and scaling, persistence, and user management are runtime concerns.

🔄

Seamless Interoperability

Use any Python library or framework. Jac maintains full compatibility with the entire Python ecosystem while adding powerful new capabilities.

Built by Nerds and Innovators

~ Imagine, Create, Launch ~

Jac is an innovative programming language with Python-like syntax that compiles to Python bytecode, JavaScript, and native machine code, providing seamless access to the PyPI, npm, and native ecosystems. Created by @marsninja with contributors of Jac Hackers Everywhere, it introduces cutting-edge programming models and abstractions specifically designed to minimize complexity and embrace AI-forward development.

Our mission is to automate categories of common software systems that typically require manual implementation, making advanced programming paradigms accessible to developers worldwide.

-- GitHub Stars

-- Forks

-- Contributors

Learn Our Story Join Community

Journey of Jaseci

~ The story of an idea realized ~

Prologue: In 2020, the concept of a new way of developing software in the AI era is concieved and later described in a Forbes blog article.
What’s In Store For The Next Generation Of AI? The Jaseci Perspective

In 2022, the idea has evolved and the first intellectual step in the journey of Jaseci and Jac is described.
The Case for a Wholistic Serverless Programming Paradigm and Full Stack Automation for AI and Beyond -- The Philosophy of Jaseci and Jac

Then in 2023, the idea survives peer-review at Computer Architecture Letters.
The Jaseci Programming Paradigm and Runtime Stack: Building Scale-out Production Applications Easy and Fast

In 2024, the idea that AI should be a conventional code construct in the language is conjoured and ellucidated.
LLMs are Meaning-Typed Code Constructs

Then in 2025, the idea survives peer-review. (pending)
Meaning-Typed Programming: Language Abstraction and Runtime for Model-Integrated Applications

That same year, "object-spatial programming" described in earlier works becomes "object-spatial programming" and is rigorously defined.
Object-Spatial Programming

And, in 2025, the notion of "scale-native programming" through langauge abstraction is rigorously defined though it was first described in the original 2022 paper.
Extending Object-Spatial Semantics for Scale Native Programming

And the journey continues...