- 1Enter your school's area
City or locality, plus school hours.
- 2We read the air
Live and forecast PM2.5, PM10 and heat, scored on India's CPCB AQI.
- 3Get a clear plan
Assembly, sports, safe hours and a parent notice in English and Hindi.
School-day air quality
Show full-day chart
Hourly values are indicative CPCB sub-indices from forecast concentrations; the official AQI uses 24-hour averages.
Which days to plan indoor activities
Ready to send
Assistant summary
No schools yet
Each morning, AWS Lambda builds each school's plan → subscribers from Amazon DynamoDB → delivered through Amazon SNS.
Delivered alerts
Locally, alerts are written to an outbox. On AWS, the same job publishes to an SNS topic (email).
About SaansSafe
A decision tool for school principals and parents on bad-air days in India.
The problem
Every winter, North Indian cities cross "Very Poor" and "Severe" air quality. Principals must decide each morning whether to hold assembly and sports outdoors, often with no clear data. Children breathe faster than adults and are among the most affected, especially those with asthma.
Who it's for
- School principals deciding on assembly, sports and outdoor time each morning.
- Parents, who get a plain notice in English and Hindi on WhatsApp.
- District officials who need to see which areas should move activities indoors.
How it works
- You enter a city or locality and school hours.
- SaansSafe fetches live and forecast PM2.5, PM10 and weather from Open-Meteo (CAMS).
- It computes India's official CPCB AQI and hourly sub-indices for the school day.
- Clear rules turn that into a plan: assembly, sports, the safest outdoor window and care for sensitive children.
- It writes a parent notice in English and Hindi to copy, share on WhatsApp or listen to.
Architecture on AWS
| AWS Lambda | Runs the FastAPI app and the daily alert job. |
|---|---|
| Amazon API Gateway | Public HTTPS endpoint for the web app and API. |
| Amazon EventBridge | Triggers the alert job every day at 7:00 AM IST. |
| Amazon DynamoDB | Stores subscribed schools. |
| Amazon SNS | Delivers the morning alert by email. |
| Strands Agents SDK | Open-source agent framework; the agent uses the same AQI and planning tools. Without a model key the app runs in deterministic mode. |
CPCB AQI scale
| AQI | Category | Possible health impact |
|---|---|---|
| 0–50 | Good | Minimal impact |
| 51–100 | Satisfactory | Minor breathing discomfort to sensitive people |
| 101–200 | Moderate | Breathing discomfort to people with lung or heart disease, children and older adults |
| 201–300 | Poor | Breathing discomfort to most people on prolonged exposure |
| 301–400 | Very Poor | Respiratory illness on prolonged exposure |
| 401–500 | Severe | Affects healthy people; serious impact on those with existing disease |
Limitations
- Forecast models can spike (e.g. dust events); hourly values are indicative, while the official AQI uses 24-hour averages.
- Model forecasts are not ground-station readings; local conditions can differ.
- Alerts on AWS currently go to every SNS email subscriber; WhatsApp delivery is planned, not live.
- The live demo runs in deterministic mode; the Strands agent layer needs a model key.
- Hindi notices come from templates and should be reviewed by a native speaker.
Data sources
- Open-Meteo Air Quality API (Copernicus CAMS forecasts)
- Open-Meteo Weather and Geocoding APIs
- CPCB National Air Quality Index breakpoints
Guidance is general and not medical advice. Follow CAQM/GRAP and state directives.