Connecting surplus food with the people who need it — powered by Jac and Jaseci.

Nourish is a full-stack autonomous system that reduces food waste by automatically coordinating surplus food from businesses with shelters and volunteer drivers. It uses Jac’s graph-native programming model to represent and solve real-world logistics as a live, traversable ecosystem.

Every day, large amounts of edible food are wasted due to overstock or expiry, while shelters struggle with limited resources.

The core issue is coordination, not supply:

• Businesses don’t know who needs food
• Shelters don’t know what is available
• Volunteers are not efficiently assigned

This creates waste and inefficiency in systems that should be connected.

Nourish solves this using an autonomous multi-agent system built on a live graph.

It automatically:

• Collects surplus food from businesses
• Matches it with shelters based on demand and capacity
• Assigns volunteers for pickup and delivery
• Dispatches everything without human intervention

Nourish uses three autonomous walkers:

• Processes incoming food donations
• Validates expiry and quantity
• Creates DonationNode

• Monitors shelter capacity and needs
• Aggregates real-time demand data

• Matches donations to shelters
• Assigns available volunteers
• Sends dispatch notifications
• Prevents duplicate processing using graph state

The entire system is modeled as a live spatial graph.

Business → Donation → Shelter
↓
Volunteer

Nodes:

• BusinessNode
• ShelterNode
• DonationNode
• VolunteerNode

Edges:

• Offers
• AllocatedTo
• AssignedVolunteer

Jac walkers traverse and modify this graph in real time.

• Jac / Jaseci — backend, AI agents, graph logic
• Jac Walkers — autonomous decision-making agents
• Jac Frontend (.cl.jac) — UI system
• REST API (Jac) — backend services
• Graph-based architecture — core system design

• Real-time donation feed
• Shelter request management
• Autonomous dispatch system
• AI agent monitoring dashboard
• Interactive map visualization
• Analytics and alerts system

• add_donation() — submit surplus food
• get_donations() — fetch all donations
• get_shelters() — shelter capacity data
• run_dispatch() — trigger autonomous cycle
• get_dashboard() — system overview
• reset_ecosystem() — reset simulation

jac install

Run backend simulation
jac run services/nourish.jac
Start full-stack app
jac start --dev main.jac

Then open:
http://localhost:8001

Project Structure
nourish/
├── main.jac
├── jac.toml
├── index.cl.jac
├── services/
│   ├── nourish.jac
│   ├── nourish_api.jac
├── pages/
├── components/
└── data/
Why Jac?

Jac is ideal for this project because:

The system is naturally a graph
Walkers act as autonomous AI agents
State is persistent and traversable
Backend, frontend, and logic are unified in one language
Perfect for real-world coordination systems
Future Improvements
LLM-based semantic matching using by llm()
Real SMS/email notifications (Twilio, SendGrid)
Live GPS volunteer tracking
Database integration (PostgreSQL / Supabase)
Mobile application
Multi-city scaling
License


Built using Jac, a graph-native programming language for autonomous systems.