Multi-Hazard Response Intelligence

From a Nokia 1100
to a satellite map
in 10 seconds

Every distress SMS becomes an AI-classified pin on a live operations map. No app. No internet. No smartphone required.

SMS → bonnya derai 40 jon help dorkar
0
Classification
0
Languages
0
Internet Required
0
People affected
2022 Sylhet flood
0
Upazilas flooded
Across 5 divisions
0
Response delay
Manual processing
0
Unified systems
Before MapSOS
The Problem

In 2022, millions were stranded with no way to call for help.

When the Surma and Kushiyara rivers breached simultaneously in Sunamganj, victims had no structured way to report their situation. Messages arrived in Bangla, English, and Banglish — via phone calls, Facebook posts, and word-of-mouth.

A food request and a drowning rescue looked identical in a coordinator's phone log. Relief boats went to accessible areas three times while cut-off zones waited.

The Solution

One SMS. AI classifies. Map updates. Help arrives.

Works on any phone — no internet, no app install, no smartphone

AI handles Bangla, English, Banglish — even with typos and missing fields

7 map layers: flood extent, shelters, vulnerability, routes, infrastructure

Two-way SMS — victim gets reply with nearest shelter and instructions

How it works

Distress call to relief
in four steps

STEP 01

Report

SMS from any phone, web form, or Bangla voice input. Works on 2G.

STEP 02

AI Classifies

Extracts hazard, location, urgency — from 3 languages with typo tolerance.

STEP 03

Map Updates

Pin drops on the live map. Coordinators see it in real-time. Toast alerts fire.

STEP 04

Help Arrives

Nearest volunteers auto-alerted. Victim gets reply SMS with shelter info.

Core Capabilities

Built for the worst conditions

SMS-First Architecture

The people who need help most have the least connectivity. MapSOS accepts distress reports via basic SMS — no internet, no smartphone, no app install. A farmer on a Nokia 1100 can report a flood the same way a coordinator views it on a satellite map.

Any 2G phone Voice input Two-way reply
SMS SIMULATION
bonnya derai 40 jon help dorkar
AI classified — FLOOD / rescue / T1
জরুরি রিপোর্ট গৃহীত। উদ্ধার দল Derai এলাকায় পাঠানো হচ্ছে। নিকটতম আশ্রয়: Derai Community Centre

AI Classification Engine

Messages arrive in Bangla, English, and Banglish — often with typos, missing fields, and wrong order. Our AI extracts hazard type, need type, location, people count, and urgency in under 10 seconds. It knows that "shatkhira" means Satkhira and "bonnya" means flood.

3 languages Typo tolerant 18 worked examples
shatkhira te bonnya 50 jon
Satkhira / flood / rescue / T1
বন্যায় ৫০ জন আটকা ধর্মপাশা
Sunamganj / flood / rescue / T1
pani lagbe 100 jon chhatak
Sunamganj / flood / water / T5

7-Layer Operations Map

Not just pins on a map. Toggle between flood extent from Sentinel-1 SAR, infrastructure exposure scoring, community vulnerability heatmaps, route accessibility, shelter capacity, and vulnerable household markers. Switch to satellite imagery for terrain context.

Satellite toggle Real-time updates 40 shelters
Reports
55+
Flood Zone
SAR
Shelters
40
Vulnerability
10 zones
Infrastructure
20
Routes
10 roads
Live Preview

See the dashboard in action

LIVE — BANGLADESH

Every message is
someone asking for help.

Open the live dashboard or submit a distress report to see the full system in action.