GDACS MCP. Track live disasters & assess impact risk.

You gotta know what’s happening when disaster strikes, globally. This server connects your AI client straight into the Global Disaster Alert and Coordination System (GDACS), letting you pull real-time hazard data without needing to mess with any API keys or complex credentials. It's a direct feed of structured world crisis info.

To track global hazards, you can get immediate updates on specific events. You’ll find get_latest_earthquakes, which shows the most recent quakes worldwide; it reports magnitude, depth, location, and even preliminary impact estimates, letting you filter by minimum size or a date range. If there are tropical cyclones out there—hurricanes or typhoons—you can use get_latest_cyclones to pull names, categories, wind speeds, locations, and the whole forecast track.

For floods, get_latest_floods lists active events globally, giving you locations, start dates, alert levels, affected population numbers, and severity scores. Volcanoes aren't exempt; get_latest_volcanoes reports current activity worldwide, listing names, locations, eruption types, alert status, and the data on ash plumes. You can track wildfires with get_latest_wildfires, which gives you fire locations, start dates, burned area metrics, alert levels, and how many people are affected.

And if drought is hitting a region, get_latest_droughts lists those events globally, tracking severity levels, alert status, and the populations impacted. For earthquakes alone, you can also run get_event_list, which searches for disaster events across types like EQ, TC, or FL, letting you filter by date range or time period to list names, types, and current alert levels.

When you need specific event details, you’ve got a few ways. You can use get_event_detail to pull full information about one specific disaster instance. If you know the ID of an immediate warning, get_alert_detail retrieves all the deep, granular details for that single alert. To check the overall status of different kinds of warnings, get_alerts gives you a list of current alerts, showing severity—red, orange, or green—and letting you filter by event type or time period.

You can also run get_event_list to search for events and get their names, types, and alert levels.

To understand the full scope of any given disaster, you calculate its impact using get_impacts. This tool provides estimates for affected population count, potential fatalities, economic losses, and confidence levels tied directly to a specific event. For mapping out exactly where the damage is, run get_event_geojson; this returns the geographic boundaries and affected areas for any specified disaster event in GeoJSON format.

You can also pull all current alerts using get_event_geojson, which helps map the boundaries of an entire disaster event.

To summarize the data retrieval process: you start by getting a list of events, then narrow down to specific details, and finally calculate the impact or visualize the area. For example, if you're dealing with a flood, you first get a general list via get_event_list, then pull its full info using get_event_detail.

Next, you can run get_impacts on that event to see the numbers and use get_event_geojson to map out the exact affected boundaries. The mechanism is built for deep data dives: if you need a single alert's history, use get_alert_detail; if you just want to know what major events are happening right now across all types, run get_alerts.

You never gotta worry about missing key data points because the server covers every facet of hazard response.

You can’t get around the specialized tracking. If you're focused purely on tropical cyclones, get_latest_cyclones handles it; if you need drought metrics, you use get_latest_droughts. The tools work together: getting a list of earthquakes via get_latest_earthquakes gives you enough info to then run get_impacts and see the projected losses for that quake.

This means your agent doesn't just check boxes; it builds out a full, actionable picture from raw global data. You'll never have to stitch together multiple APIs or worry about incompatible data formats because this whole system is designed to feed structured hazard information directly into your workflow.