This one is actually in pretty good shape — no track names are hardcoded anywhere so the Dev Tooling / Open Innovation removal doesn't affect it. Just two small things worth updating:

  1. The intro says "Jac was created by Professor Jason Mars at the University of Michigan" — worth adding his title as founder of Jaseci Labs since that's the organization running the event.
  2. "Email: Contact the organizing team through the website" — if you want, you can just put svatsal@umich.edu directly here so people don't have to hunt for it.

Here's the clean version:

Introduction

JacHacks is built entirely around Jac and Jaseci, the world's first agentic programming stack. Jac was created by Professor Jason Mars at the University of Michigan. Jason is a CS professor at Michigan, founder of Jaseci Labs, founder of Clinc (an AI company that was acquired), and one of the leading researchers in agentic AI systems. He and his team have been building Jac as a fundamentally new way to develop AI applications — one where the language itself understands agents, graphs, and LLM-powered functions as first-class citizens.

The resources below will help you go from zero to building with Jac, whether you're a complete beginner or an experienced developer looking to understand the agentic programming model.

Tools and Technologies

  • Jac Language — The core programming language you'll be building with. Supersets Python and JavaScript. Install with pip install jaclang. [Jac Documentation]
  • Jaseci — The full-stack runtime and platform that powers Jac applications. Handles deployment, scaling, and infrastructure automatically
  • by llm() — Jac's signature feature. Replace entire function bodies with a single keyword. Your type annotations become the prompt. No prompt engineering required
  • Walkers and Graphs — Jac's native abstractions for building multi-agent systems. Walkers traverse graphs to model complex agentic workflows without external orchestration libraries
  • jac start — One command to go from a single .jac file to a fully deployed application with MongoDB, Redis, JWT auth, and Kubernetes support

Inspiration

  • Build an autonomous research agent that crawls papers, summarizes findings, and maintains a knowledge graph
  • Create an AI-powered code review tool that understands entire codebases through graph traversal
  • Design a multi-agent trading system where specialized agents handle different aspects of market analysis
  • Build a personal AI assistant that manages your calendar, email, and tasks through conversational walkers
  • Create a collaborative writing tool where multiple AI agents each handle different aspects of content creation
  • Build a healthcare triage tool, a climate data analyzer, or an accessibility platform powered by Jaseci agents

These are just starting points. The best projects at JacHacks will be ones nobody has thought of yet.

Contact & Support

  • Website: jachacks.org
  • Email: svatsal@umich.edu
  • Discord: Join the JacHacks Discord for pre-event team formation, Q&A, and live support during the hackathon (link on website)
  • During the event: Jaseci team members and mentors will be on the floor throughout the hackathon for hands-on technical help with Jac

 

Additional Resource Links