Inspiration

This is IPC Nexus, built by Team Zebulun. Hunger doesn't arrive overnight. The warning signs show up months earlier, but they sit inside long reports that busy people rarely have time to read. So action often starts late, when families are already in crisis. We wanted to close that gap."

What it does

"The idea is simple: see hunger coming, and act before it becomes famine.

what it does

"This is one screen covering 45 countries and regions monitored by the IPC and the Cadre Harmonisé. Each color is a food insecurity phase, from minimal to famine. Instead of digging through a report, you can see where the pressure is in a few seconds."

"Let me tap North Darfur. Here I can see the current phase, the people affected, and a data coverage check that shows how much of the population was actually analyzed. We added that on purpose, because a number without context can mislead people. Down here is the sub-national table, so you can go from a country down to a locality."

"Now the part I like most. The shock simulator. Say supply corridors are cut for thirty days, or food prices rise, or conflict gets worse. I move these controls and the risk picture shifts straight away. This is an illustrative scenario, not a forecast. It helps someone ask 'what if' before decisions are made." "Next is Ask. I type Lesotho, and it builds a brief from the IPC data loaded into the app: phase numbers, the units inside the country, and the limits of the data. Look at the suggested actions. There is no escalation action, because Lesotho has no Phase 4 population." "Now Sudan. This one does show an escalation action, because the data has Phase 5 populations. Every action names the figure that triggered it. To be clear about how this works: Ask does not use an AI model, and it does not invent numbers. It matches your search to the data and builds the brief from it."

"And when I search somewhere the IPC does not monitor, like Germany, the app says there is no active classification. It used to show North Darfur as an emergency. More on that in a moment."

How we built it

"Now, how we built it. We had never used Manus before this hackathon, and it was our first time adding a map to a project. So we didn't start by prompting. We wrote six documents first: product requirements, technical requirements, a UI and UX brief, the app flow, the backend schema with the real IPC data, and an implementation plan. Writing them first was the biggest lesson. When our instructions were clear, the tool built what we meant."

Challenges we ran into

"Three things went wrong. First, the map. Halfway through, a watermark appeared saying an API key was required. We swapped to keyless Esri and Humanitarian OpenStreetMap tiles. Second, the Germany bug. A leftover fallback in the code showed the first zone whenever a search had no match. In a food security tool, the wrong crisis for the wrong place is dangerous, so we removed it. Third, some labels claimed too much, so we renamed the Co-Pilot and made every action depend on the data.

Accomplishments that we're proud of

The lesson: test the edges, not just the happy path."

What we learned

we learnt to you maps, learned to resolve bug problem and the we enjoyed using manus

What's next for IP NEXUS

"IPC Nexus is decision support, not an official famine classification, and the app says so. There are a few known issues we list openly in the README. Next, we will fix those and start our fintech project for the next hackathon, writing the specs first again. Thank you for watching."