Haze Day
Live haze readings and what to do about them.
Live PM2.5 readings and plain-language guidance for Southeast Asia — official NEA data for Singapore, labelled model estimates for 79 places, an AI explainer and a Telegram alert bot.
> cat ./overview.md
Haze Day (烟霾助手) answers the question every parent, coach and runner asks during haze season: can we go outside right now? Each page leads with the live 1-hour PM2.5, the NEA band it falls in, and what that means for sport, school PE, commuting, laundry, pets and the home.
Singapore readings come straight from NEA via data.gov.sg; 78 other cities across Malaysia, Indonesia, Thailand, Vietnam, the Philippines and the rest of ASEAN show CAMS model estimates, always labelled as non-official. A live map layers the PM2.5 field, NASA FIRMS fire hotspots, 1.5 km wind and rain radar, and a "why is it hazy?" panel traces upwind fire regions.
AI only explains data that already exists — it never produces readings. It writes the plain-language summary on each page, compares haze headlines across Singapore, Malaysian and Indonesian outlets, and answers questions in the @hazeday_bot Telegram bot, which also sends alerts at a cut-off you choose.
> ls ./features
Official + Modelled, Clearly Labelled
NEA 1-hour PM2.5 and 24-hour PSI for Singapore; CAMS estimates elsewhere, never presented as official.
Situation Guides
Team sport, running, school PE, commute, windows, laundry, outdoor events and pets — each mapped to NEA bands and the 110 µg/m³ youth-league cut-off.
Live Haze Map
PM2.5 field, fire hotspots, upper-level wind and rain radar with a −12 h to +24 h time slider.
Why Is It Hazy?
Upwind analysis against hotspots in Riau/Jambi, South Sumatra and Kalimantan from NASA FIRMS satellite passes.
Telegram Alert Bot
@hazeday_bot sends alerts at your cut-off, supports a sensitive-person mode and answers questions from the live data.
Open Data & Widget
Public JSON API, a cookie-free embeddable widget and llms.txt for AI tools — CC BY with attribution.
> open ./screenshots
> cat ./stack.json
{
"name": "Haze Day",
"category": "Public Health / Air Quality",
"stack": [
"NEA / data.gov.sg",
"NASA FIRMS",
"CAMS · Open-Meteo",
"Leaflet",
"Telegram Bot",
"Next.js"
]
}
> grep -l "Haze Day" ./blog
Haze Day: Official Data First, AI Only to Explain
How we built a haze tracker that parents, coaches and runners can act on: the 1-hour PM2.5 up front, model estimates always labelled, and an AI that never invents a reading.
Nine Products, One Stack: The NGMOB Lineup
Haze Day, Hanakoi, TokenLiquid and Doubo join Floniks, Kolens, GoGlow, Naga and Hollow Lullaby. Here is what each one does, and where MCP, agents and orchestration fit in.