For Newsrooms & journalists
A local breaking-news wire for newsrooms
Local news desks used to monitor scanners; now the signals are scattered across 689 official feeds and thousands of local-media reports. CityAlert normalizes all of it — police portals, fire and wildfire services, weather and disaster agencies, plus geocoded local-media reporting in 22 languages — into one stream. When a region's incident rate jumps against its own 30-day baseline, the trending surface flags it: that's often a story breaking before the wires pick it up.
- feeds polled
- 689
- focus regions
- 757
- live map
- free
What you get
Trending areas — stories as they break
A statistical detector compares each region's live incident rate against its own 30-day baseline and flags anomalies. When a city's rate doubles, there's usually a reason — a fire, a storm, a police operation — and you see it on one screen instead of a dozen feeds.
The media firehose, geocoded
Local media in 22 languages is scraped, translated and geocoded into incidents — the same signal newsrooms watch for, already normalized with source links back to the original reporting. Buenos Aires, Istanbul, Athens and Hanoi are among the highest-volume streams.
Per-city Atom feeds & free API
Every one of the 157 city pages publishes an Atom feed of its latest incidents, and the live GeoJSON API needs no key for real-time reads. Wire it into your newsroom dashboard, Slack or alerting pipeline in an afternoon.
Embeddable incident widgets
Drop a live incident summary into a story or a standing 'what's happening' page with one iframe — scoped to a city or region, refreshing every ~10 minutes, linking back to the full map.
Community reports as a tip line
Residents submit what they see — confirm/resolve voting filters the noise. For newsrooms covering covered regions, it's a public tip stream you don't have to moderate yourself.
Historical archive for context
Every incident ever ingested is archived — 'how unusual is this?' becomes a query, not a guess. Crime year-over-year and event-history endpoints back the numbers behind the story.
What it won't do
We'd rather lose a deal than oversell the data — these are the honest limits for newsrooms & journalists.
No on-scene media
Incidents are data points from official feeds and geocoded reports — no photos, video or audio. For verified scene media, your own reporters and tools like Citizen remain the source.
Media-derived incidents are approximate
Stories are geocoded to block or neighbourhood level and labelled as such. They're leads to verify, not publishable fact — always confirm with the linked source before running.
Coverage follows the feeds
Depth is strongest in the 757 defined focus regions. Elsewhere, global hazard feeds (earthquakes, disasters, weather) still work — but don't expect incident-level crime data where no feed exists.
Is it a fit?
Built for you if…
- Local and regional newsrooms that want a single incident wire instead of monitoring dozens of feeds.
- Breaking-news desks that need early signal on events escalating in their coverage area.
- Data journalists who want normalized, queryable incident history for context and trends.
- Publishers who want a live incident widget on their site without building the pipeline.
Probably not if…
- You need verified on-scene photos and video — CityAlert carries data points, not media.
- You want to publish incidents verbatim without checking the linked source. These are leads, not wire copy.
Free for most desks; API tiers for volume
The map, trending surface, Atom feeds, embeds and the real-time API are free with attribution. Newsrooms that need the full history archive, higher request volumes or push-webhook delivery use the paid API tiers — self-serve keys, no sales call.
Full plan details on the pricing page; questions and enterprise scope: [email protected].
Frequently asked questions
Can we republish CityAlert data?
Yes — our normalization layer is CC BY 4.0, so cite CityAlert and the underlying source named on each incident. The upstream feeds carry their own licenses (listed on /sources); official government feeds are generally fine, media-derived incidents should be treated as leads pointing back to the original reporting.
How fresh is the data?
Feeds poll on their own cadence — most land within ~10–15 minutes of upstream publication. It's fast enough to tip a story, not fast enough to replace monitoring an active scene.
Can we get alerts pushed to us instead of polling?
Yes — paid plans support outbound webhooks with category, severity, region and geo-fence filters, so a newsroom Slack or CMS can get incidents delivered rather than scraped.
CityAlert.live is an aggregation and awareness tool — media-sourced incidents are approximate and every pin names its source. For emergencies always follow official channels. See our terms and sources & licenses.