Skip to main content
Glama
nanzz-leo

DRIP

by nanzz-leo
README.md
#  DRIP – Disaster Response Intelligence Platform

> AI-powered emergency coordination platform built with NitroStack MCP to accelerate disaster response through intelligent planning, resource orchestration, and real-time operational insights.

##  Overview

DRIP (Disaster Response Intelligence Platform) is an AI-driven emergency response system designed to assist Disaster Management Authorities, Emergency Operations Centers (EOCs), first responders, and relief organizations.

Instead of relying on fragmented information and manual coordination, DRIP orchestrates specialized intelligence modules to:

- Assess disaster severity
- Analyze environmental conditions
- Locate critical resources
- Coordinate rescue operations
- Generate actionable Situation Reports for decision makers

Built using **NitroStack MCP**, DRIP exposes intelligent tools that can be orchestrated by AI agents to support emergency response workflows.

---

# Features

## AI Planner

- Intelligent incident analysis
- Dynamic task selection
- LLM-powered Situation Report generation
- Minimal task execution strategy
- Hallucination-resistant reporting

---

## Intelligence Module

Provides:

- Weather assessment
- Hazard analysis
- Safe route identification
- Nearest hospital discovery
- Shelter discovery
- Severity estimation

---

## Operations Module

Coordinates operational response:

- Incident reporting
- Rescue vehicle allocation
- Volunteer assignment
- Shelter updates
- Relief inventory management
- Emergency deployment workflow

---

## Situation Reports

Automatically generates structured operational reports including:

- Executive Summary
- Severity Assessment
- Affected Area
- Threat Analysis
- Resource Status
- Tactical Action Plan
- Intelligence Gaps
- Operational Rationale

---

# Architecture

```
                    User Incident
                          │
                          ▼
                 Planner Orchestrator
                          │
          ┌───────────────┼───────────────┐
          ▼               ▼               ▼
   Intelligence     Resources      Operations
          │               │               │
          └───────────────┼───────────────┘
                          ▼
               AI Situation Report
```

---

# Built With

- NitroStack MCP
- TypeScript
- Node.js
- NitroCloud
- Gemini API
- Zod
- Mapbox (Widgets)

---

# Project Structure

```
src
│
├── planner/
│   ├── planner.module.ts
│   ├── planner.service.ts
│   └── planner.tasks.ts
│
├── modules/
│   ├── intelligence/
│   ├── operations/
│   ├── resources/
│   └── pizzaz/
│
├── prompts/
│   ├── systemPrompt.ts
│   └── responsePrompt.ts
│
├── services/
│   └── llm.ts
│
└── app.module.ts
```

---

# Running Locally

```bash
git clone https://github.com/nanzz-leo/DRIP.git

cd DRIP

npm install

npm run build

npm run dev
```

---

# Example Workflow

```
Incident Report
        │
        ▼
Planner analyzes incident
        │
        ▼
Assess Disaster
        │
        ▼
Locate Resources
        │
        ▼
Plan Rescue
        │
        ▼
Generate AI Situation Report
```

---

# Design Principles

- Human life first
- Verified information only
- Zero hallucination policy
- Modular AI orchestration
- MCP-native architecture
- Extensible task-based design

---

# Team

Built during the **NitroStack × Amrita University Hackathon**.

TDQS

B3.4/5.0

Scored across 4 tools

Disambiguation5/5

Each tool has a clearly distinct purpose: locating hospitals, locating shelters, finding safe routes, and geocoding. No two tools overlap in function; even the two 'nearest' tools target different types of facilities.

Naming Consistency5/5

All tool names follow a consistent verb_noun pattern using camelCase (findNearestX, geocodeLocation). The naming is predictable and readable.

Tool Count5/5

Four tools is a reasonable number for a focused emergency response server. Each tool earns its place without being excessive or insufficient.

Completeness4/5

The tool set covers the core emergency location needs: finding hospitals, shelters, safe routes, and geocoding. Minor gaps may exist (e.g., getting detailed info on a facility), but the surface is largely complete for its stated purpose.

Maintenance

ActivityStale
ResponsivenessNo issues