NAGAR AI

Ward Control Room · Chennai Corp Demo Zone

Citizens affected— Open—
IST--:--:--
LIVE

File a complaint

Voice · Photo · Text — any language

TKT-0000
Tamil, Hindi, English or mixed — the extractor reads all of it.
Location auto-detects when allowed — or type the area name below.
Category—
Severity—
Location resolved—
Clean summary
Photo metadata (EXIF)
Dedup result—

Live intake log

Track your complaint

Enter the complaint ID you received when filing

TKT
Shown on the receipt after you file — a short number like #42.
—
Open clusters
—
Citizens affected
—
Reports auto-merged
—
SLA breaches

Operations board

— clusters

PRT formula

Regulation · R-7
PRT = severity(1–5) × √(affected_citizens) × min(days_pending, 30) × proximity_multiplier
proximity_multiplier = 1.5 if any report in the cluster is within 200 m of a geocoded school/hospital, else 1.0.
√(affected) dampens brigading — 10 duplicate reports gives ×3.16, not ×10. days_pending is capped at 30 so an old, forgotten complaint can't mathematically outrank an active emergency forever. Severity and location must still be independently corroborated by ≥ 1 report before the multiplier applies, so a single fake "near hospital" tag can't dominate the score on its own.
Worked example loads after you run the judging set below ↓

Judging set · live dedup

Similarity · text / geo / vision