USGS Recent Earthquakes
FreeGET /api/climate/earthquakesRecent earthquakes from the USGS Earthquake Hazards Program pre-built summary feeds. The upstream feeds refresh every minute. TensorFeed proxies them with a per-(magnitude, period) KV cache whose TTL scales with the feed window (60s for the hour feeds, 900s for the month feeds). The response flattens the upstream GeoJSON FeatureCollection into a plain earthquakes list so most agents can consume it without geometry-processing knowledge.
When to use this endpoint
When an agent needs the most recent earthquake activity globally — for situational-awareness briefs, climate/disaster reporting, alert routing for tsunami-flagged events, or pairing with population/news feeds. Choose the magnitude bucket (significant means 'noteworthy by USGS criteria,' the numeric buckets are M-or-above thresholds, 'all' returns every detected event) and the time window (hour, day, week, month).
Parameters
| Name | In | Type | Description |
|---|---|---|---|
| magnitude | query | string | significant | 4.5 | 2.5 | 1.0 | alle.g. 4.5 |
| period | query | string | hour | day | week | monthe.g. day |
| limit | query | number | Max earthquakes to return (1-500)e.g. 50 |
* required
Example response
{
"ok": true,
"tier": "free",
"source": "live",
"fetched_at": "2026-05-09T17:08:00Z",
"query": { "magnitude": "4.5", "period": "day", "limit": 50 },
"feed_metadata": {
"title": "USGS Magnitude 4.5+ Earthquakes, Past Day",
"generated": "2026-05-09T17:07:42Z",
"api_version": "1.10.3",
"upstream_count": 12
},
"count": 12,
"earthquakes": [
{
"id": "us7000abc",
"magnitude": 5.1,
"magnitude_type": "mb",
"place": "47 km NE of Tinogasta, Argentina",
"title": "M 5.1 - 47 km NE of Tinogasta, Argentina",
"time": "2026-05-09T15:30:00Z",
"updated": "2026-05-09T15:31:40Z",
"depth_km": 165,
"longitude": -67.55,
"latitude": -28.06,
"tsunami": false,
"significance": 401,
"felt": 12,
"alert": "green",
"status": "reviewed",
"url": "https://earthquake.usgs.gov/earthquakes/eventpage/us7000abc"
}
],
"attribution": {
"source": "USGS Earthquake Hazards Program — Real-time Feed",
"license": "US Government work in the public domain (17 USC §105)",
"redistribution": "commercial-permitted"
},
"allowed": {
"magnitudes": ["significant", "4.5", "2.5", "1.0", "all"],
"periods": ["hour", "day", "week", "month"]
}
}Code samples
Python SDK
from tensorfeed import TensorFeed
tf = TensorFeed()
data = tf._get("/climate/earthquakes", magnitude="4.5", period="day")
for q in data["earthquakes"][:10]:
print(f"M{q['magnitude']:.1f} {q['place']} ({q['time']})")TypeScript SDK
const res = await fetch("https://tensorfeed.ai/api/climate/earthquakes?magnitude=significant&period=week");
const { earthquakes } = await res.json();
for (const q of earthquakes) console.log(`M${q.magnitude} ${q.place}`);MCP tool
Available via the TensorFeed MCP server as get_recent_earthquakes. Add npx -y @tensorfeed/mcp-server to your Claude Desktop or Claude Code MCP config.
FAQ
What does each magnitude bucket mean?
"significant" is USGS's curated stream of noteworthy events using internal criteria (large magnitude, populated area, tsunami flag, etc). The numeric buckets ("4.5", "2.5", "1.0") are simple magnitude-or-above thresholds. "all" returns every detected event in the window, including microquakes.
How fresh is the data?
USGS regenerates each summary feed once per minute. TensorFeed caches each (magnitude, period) tuple in KV with a TTL scaled to the feed window — 60 seconds for the hour feeds, 120 seconds for the day feeds, 300 seconds for the week feeds, 900 seconds for the month feeds.
What is the tsunami flag?
USGS sets tsunami = true when an event meets the criteria for tsunami evaluation (typically large magnitude in or near oceanic regions). It does not mean a tsunami was observed; it means the event is in the queue to be evaluated. For confirmed tsunami status, follow the USGS detail URL for that event.
License?
USGS data is US Government work in the public domain (17 USC §105). Commercial redistribution is permitted; the standard attribution block ships on every TensorFeed response.