How Floodi works

Live source status

Reflects this browser session's own connections to each source, refreshed automatically. "Unavailable" means the last fetch attempt failed -- not that the underlying government service is down for everyone.

Ingest pipeline health

Floodi has no separate cron-execution log to check independently -- pg_cron's own run history isn't exposed to the public site, and shouldn't be. This instead reports how fresh the newest row in Floodi's own database is, which is the closest honest proxy for "is the scheduled poller still running." A gap of more than about 15 minutes here means the poller likely isn't running, not just that nothing new has happened.

Right now

"67 counties" is Florida's real, fixed county count -- the number above is a live count of counties with an actual open NWS alert right now, not a static claim. A county reading 0 means no active alert exists for it at this moment, not that Floodi has no coverage there: USGS/NOAA gauge and FEMA zone data are queried statewide regardless of county.

Community reports

The pipeline, in plain language

1. 1. Official alerts (NWS). Flood, flash-flood, coastal-flood, hurricane, and tsunami alerts come directly from the National Weather Service's public API, polled every 2-3 minutes and pushed to your browser instantly when Floodi's own backend detects something new.
2. 2. River & stream gauges (USGS + NOAA). Real-time water levels come from USGS. A flood category (Action, Minor, Moderate, or Major) is shown only when NOAA has published that category for that gauge. A gauge without one shows the measured level and “No official flood stage for this gauge.”
3. 3. Coastal water levels (NOAA CO-OPS). Tide stations show the measured water level and how it compares with the predicted tide. Floodi does not assign a surge category from that difference.
4. 4. Flood zones (FEMA NFHL). Designated Special Flood Hazard Areas, queried live from FEMA's own map service for whatever area you're viewing.
5. 5. Community reports. Residents submit what they're seeing, with an optional photo. Reports get an AI photo-moderation pass (checks the photo actually shows relevant conditions and contains nothing unsafe to publish) before any photo is shown publicly.
6. 6. Confirmation voting & time decay. Other nearby users can confirm or dispute a report. A report with more disputes than confirmations, or one that's aged past its expected relevance window without new confirmations, is treated as lower-confidence and eventually marked cleared -- Floodi doesn't keep showing stale reports indefinitely.

Honest limitations

• Community reports are unverified by default -- confirmation counts and admin verification raise confidence, but Floodi cannot guarantee accuracy of a self-reported observation.
• A gauge without an official NOAA flood-category threshold is shown as a measurement only. Floodi does not assign Action, Minor, Moderate, or Major from a generic water level.
• FEMA flood zones reflect the National Flood Hazard Layer as FEMA currently publishes it, which can lag real-world conditions after a map revision.
• Every "live" status above depends on this browser successfully reaching each government API -- a source reading "unavailable" here reflects this session, not necessarily every visitor's.
• Storm-surge figures shown on property-risk pages are Floodi's own latitude/FEMA-zone estimate, not official NOAA NHC SLOSH model output.