1. 1Our monitors watch official pages, local news and public social feeds, and notice a change within seconds. Watching costs nothing.
2. 2Every change is captured through Context.dev: a fresh scrape plus structured extraction of each claim, with a word-for-word quote as evidence.
3. 3An LLM (grouping model) groups claims that say the same thing across sources. It never decides what is true.
4. 4Transparent rules set each claim's status: verified only by an official source, corroborated by independent origins, false when an official denies it, risky reassurances held back.
5. 5The board updates live. Context.dev's own semantic monitors run in parallel as a second opinion on every official page change.
6. 6Every change is kept as a time series in Tiger Data (how long a rumour stood before officials debunked it), and the key facts can be heard as a spoken bulletin voiced by ElevenLabs, in English and the local language.

The scenario towns, agencies and posts are fictional; the yellow/orange/red weather warnings are real FMI data.