Welcome

Integrate the most accurate hyperlocal weather data

Building a commercial application or GIS workflow with weather data? Book a live demo to talk with our team.

Try our API now

Why use the Precip API?

Rainfall and surface weather vary sharply over short distances. One site can get two inches of rain while another ten minutes down the road stays dry.

Most weather APIs serve coarse model forecasts and relabel them as history. Precip runs a closed-loop observation and forecast stack that is benchmarked in public against NOAA stations:

1 km observed rainfall and snowfall history (2019 to present): Built from 180+ NEXRAD Doppler radar stations and quality-controlled rain gauges. Precip is 60% more accurate than a NOAA HRRR forecast at recording what actually fell, with 275% better rainfall detection (see the history benchmark).
1 km next-hour rain nowcast: Updated every 5 minutes, with 1-minute precipitation rates via /api/v1/minutely. Ranked #1 in 37 of 38 live accuracy tests against Apple Weather, Google Weather, Tomorrow.io, and Rainbow.ai at NOAA climate reference gauges (see the nowcast benchmark).
2.5 km 10-day hourly forecast: Neural downscaling on Google DeepMind's WeatherNext 3 plus a multi-model precipitation blend (WeatherNext 3, ECMWF IFS ensemble, and NOAA HRRR). Ranked #1 in 72 of 75 benchmark tests against ECMWF IFS HRES and raw WeatherNext 3 across ~2,200 U.S. airport weather stations (see the 10-day forecast benchmark).
Soil, snow, and surface weather: Query root-zone soil moisture and soil temperature (0–10 cm, 10–40 cm), snowpack depth, 2 m air temperature, humidity, wind speed and gusts, pressure, cloud cover, and solar radiation.

Five ways to integrate

REST Point API: Query hourly, daily, minutely, and rolling-window data in JSON, GeoJSON, or CSV.
AI Assistants & Agents (MCP): Connect Claude, ChatGPT, Cursor, or custom agents to https://api.precip.ai/mcp via OAuth or Bearer API key.
Pay-per-use (x402): Query https://api.precip.ai/x402/... or https://api.precip.ai/mcp-x402 with zero signup and pay per lookup in USDC.
ArcGIS ImageServer & Web Map Tiles: Add time-enabled Esri ImageServer layers or XYZ raster tiles to ArcGIS Pro, ArcGIS Online, QGIS, Mapbox GL, MapLibre, or Leaflet.
Embeddable Weather UI: Drop a responsive, themeable weather dashboard into your web app with a single <iframe> (/embed/location).

Coverage

Data coverage and layer specifications are documented on our data availability page →

Pricing and access tiers

We offer four ways to access the Precip API:

Free ($0): Any Precip account can make 25 weather lookups a month through an AI assistant (MCP).
Personal (included with any Precip app plan): Rate-limited for home automation and personal projects (100 requests/day, 1 location per request). See rate limits and Home Assistant.
Pay-per-use (from $0.002 per lookup): No subscription or account required. Pay per request with stablecoins via the x402 protocol →
Commercial: Higher rate limits, multi-location batching, bulk spreadsheet export, map tiles, and SLAs. Book a demo →

Legal

Usage of the Precip API is subject to our Terms of Service.

Contact

For questions, comments, or feedback, email [email protected].

Last modified on October 31, 2026

Getting Started

Why use the Precip API?

Rainfall and surface weather vary sharply over short distances. One site can get two inches of rain while another ten minutes down the road stays dry.

Most weather APIs serve coarse model forecasts and relabel them as history. Precip runs a closed-loop observation and forecast stack that is benchmarked in public against NOAA stations:

1 km observed rainfall and snowfall history (2019 to present): Built from 180+ NEXRAD Doppler radar stations and quality-controlled rain gauges. Precip is 60% more accurate than a NOAA HRRR forecast at recording what actually fell, with 275% better rainfall detection (see the history benchmark).

1 km next-hour rain nowcast: Updated every 5 minutes, with 1-minute precipitation rates via /api/v1/minutely. Ranked #1 in 37 of 38 live accuracy tests against Apple Weather, Google Weather, Tomorrow.io, and Rainbow.ai at NOAA climate reference gauges (see the nowcast benchmark).

2.5 km 10-day hourly forecast: Neural downscaling on Google DeepMind's WeatherNext 3 plus a multi-model precipitation blend (WeatherNext 3, ECMWF IFS ensemble, and NOAA HRRR). Ranked #1 in 72 of 75 benchmark tests against ECMWF IFS HRES and raw WeatherNext 3 across ~2,200 U.S. airport weather stations (see the 10-day forecast benchmark).

Soil, snow, and surface weather: Query root-zone soil moisture and soil temperature (0–10 cm, 10–40 cm), snowpack depth, 2 m air temperature, humidity, wind speed and gusts, pressure, cloud cover, and solar radiation.

Five ways to integrate

REST Point API: Query hourly, daily, minutely, and rolling-window data in JSON, GeoJSON, or CSV.

AI Assistants & Agents (MCP): Connect Claude, ChatGPT, Cursor, or custom agents to https://api.precip.ai/mcp via OAuth or Bearer API key.

Pay-per-use (x402): Query https://api.precip.ai/x402/... or https://api.precip.ai/mcp-x402 with zero signup and pay per lookup in USDC.

ArcGIS ImageServer & Web Map Tiles: Add time-enabled Esri ImageServer layers or XYZ raster tiles to ArcGIS Pro, ArcGIS Online, QGIS, Mapbox GL, MapLibre, or Leaflet.

Embeddable Weather UI: Drop a responsive, themeable weather dashboard into your web app with a single <iframe> (/embed/location).

Pricing and access tiers

We offer four ways to access the Precip API:

Free ($0): Any Precip account can make 25 weather lookups a month through an AI assistant (MCP).

Personal (included with any Precip app plan): Rate-limited for home automation and personal projects (100 requests/day, 1 location per request). See rate limits and Home Assistant.

Pay-per-use (from $0.002 per lookup): No subscription or account required. Pay per request with stablecoins via the x402 protocol →

Commercial: Higher rate limits, multi-location batching, bulk spreadsheet export, map tiles, and SLAs. Book a demo →