I am looking for a weather API that provides historical hail data for insurance claims analysis.

Yes. For insurance claims analysis, I’d focus on APIs/data services that provide historical hail footprints with geographic geometry and hail severity, rather than ordinary weather-history APIs.

Strong options

• weatherguidance.com — particularly relevant to insurance. It provides historical hail-swath polygons, maximum hail size, date/time windows, and geospatial formats such as JSON/GeoJSON. Its documented hail history goes back to August 2009. It explicitly supports claims, CAT response, underwriting, and portfolio analytics.WeatherGuidance
• swathapi.com — a newer REST API built around NOAA MRMS radar-verified hail. It can return storm swaths and, importantly, property-level exposure with measured hail size. Its storm endpoint supports querying historical verified events by bounding box, time window, and hail type.Swath API+1
• explore.disasteraware.com — provides event-level historical catastrophe data, including hail swaths, tornado paths and other perils, intended for insurance, underwriting, claims, and risk analytics.DisasterAWARE
• NOAA/NCEI + MRMS — the public-data route. NOAA's MRMS system provides radar-derived products, while NCEI provides historical weather data through APIs. This can be substantially more work because you'll generally need to process/derive the hail exposure yourself rather than getting a claims-ready hail swath API.VLab+1

What I'd look for in your use case

For a claims-analysis pipeline, the useful schema is something like:

claim_id
loss_date
loss_lat
loss_lon
hail_event_id
hail_max_size
distance_to_swath
inside_swath
swath_geometry
event_start
event_end
data_source

Then you can answer questions such as:

"Was this insured property inside a verified hail swath on the reported date of loss, and what was the maximum estimated/observed hail size at that location?" StormIntel looks particularly aligned with that workflow because it explicitly supports historical validation against a book of business and provides hail polygons with severity and dates.WeatherGuidance

If you tell me (1) U.S. only or worldwide, (2) how many years of history you need, and (3) whether you need address-level results or just lat/lon + swath polygons, I can narrow this down to 2–3 APIs and compare coverage, historical depth, pricing, API format, and accuracy.

Yes. For insurance claims analysis, I’d separate the options into three categories: raw government observations, radar-derived hail footprints, and insurance-focused commercial APIs.

API / dataset	Historical hail	Granularity	Best use
NOAA SWDI / NCEI	Yes	Radar detections + reports	Building your own claims-analysis pipeline
NOAA Storm Events Database	Yes, back to 1950	County/event level	Historical claim/event corroboration
Xweather Hail Archive	Yes	Geographic queries, hail size/intensity	Convenient commercial API
Swath API	Yes/recent historical coverage	Radar-verified swaths + properties	Property-level claims/impact analysis
HailTrace API	Yes	Location/property impact history	Insurance/roofing workflows
Half Mile Hail API	Yes	~0.1° grid tile	Simple, inexpensive screening

1. NOAA SWDI — strong foundation for your own system

data.noaa.gov

NOAA's Severe Weather Data Inventory (SWDI) exposes NEXRAD Level-III hail signatures, storm cells, and preliminary storm reports, with downloadable CSV, XML, KMZ and Shapefile formats.NOAA Data+1

This is particularly useful if you want to take:

claim address → lat/lon → historical hail events → distance/time correlation → claim validation

The main drawback is that you'll have to build more of the geospatial processing yourself.

2. NOAA Storm Events Database — excellent historical corroboration

ncei.noaa.gov

NOAA's Storm Events Database contains hail events going back decades, with records beginning in 1950. Hail records are generally organized at the county/event level, rather than providing a precise hail footprint at an individual property.NCEI+1

I'd use this as a secondary source, especially for establishing that a reported hail event occurred in the vicinity of a claim.

3. Xweather Hail Archive — convenient commercial API

xweather.com

Xweather has a dedicated historical /hail/archive endpoint providing historical hail occurrences, including hail size and intensity, queried by coordinates/location. Its current archive starts in 2024 for the U.S.Xweather

This is attractive if you want a normal commercial JSON API rather than assembling NOAA datasets yourself.

4. Swath API — worth looking at for property-level insurance analysis

swathapi.com

Swath says it derives hail from NOAA MRMS radar and provides radar-verified storm swaths that can be joined to parcel/property data. Its API can return properties affected by a storm rather than merely telling you that hail occurred somewhere nearby.Swath API

That model is particularly relevant if your analysis looks like:

claim → property → hail swath → maximum expected hail size → distance/time → supporting evidence

5. HailTrace — insurance-specific workflow

developers.hailtrace.com

HailTrace exposes location impact-history functionality and explicitly supports insurance-related fields such as carrier, claim number and roof replacement date.HailTrace API

I'd investigate this if you need claim-oriented reports rather than raw meteorological data.

6. Half Mile Hail — simple historical lookup

halfmilehail.com

This is a lightweight option that returns historical NOAA NEXRAD hail detections for a U.S. address or lat/lon. It reports results for approximately a 0.1° grid tile, so it explicitly should not be interpreted as proof that hail struck an individual roof.Half Mile Hail

What I'd choose for an insurance claims system

If you're building this into an actual claims analytics/validation product, I'd lean toward:

NOAA SWDI/MRMS → geospatial processing → your claims database

rather than relying exclusively on a point-based "hail happened here" API.

The important distinction is that a hail report, a radar hail detection, a radar-derived hail-size estimate, and actual physical roof damage are four different things. For defensible claims analysis, retaining the underlying event geometry, timestamp, hail-size measurement/estimate, data source, and uncertainty is valuable.

If you tell me your requirements—e.g. U.S. only vs. global, historical depth (5/10/30+ years), address-level vs. polygon/swath data, expected API budget, and whether you need hail size—I can narrow this to 2–3 APIs and compare their endpoints, coverage, pricing, and example requests.

For insurance claims analysis, you’ll usually want more than a general historical weather API. Look for data that provides hail occurrence, hail size/severity, exact time, geospatial footprint, and evidence suitable for claim validation. Good options include:

API / Data source	Best fit	Historical hail capability	Notes
Xweather	Claims validation and property-level workflows	Dedicated hail archive with hail size, severity, timestamps, and location queries	Strong fit if you need a ready-to-use hail endpoint. The archive provides hail event data and is designed for hail risk and claims use cases.Xweather+1
NOAA radar datasets (NEXRAD/MRMS)	Building your own analytics model	Long historical record of radar-derived hail indicators such as MESH/POSH	Best for insurers with data science teams; requires more processing.Weather Developer Portal
Verisk Benchmark API	Enterprise claims workflows	Property-specific historical hail, wind, lightning, and other peril verification	Designed specifically for claims and underwriting use cases.Verisk Integration+1
The Weather Company HailVision	High-resolution hail mapping	Radar-derived hail size estimates (MESH) and hail visualization/vector products	Useful for spatial hail swaths and portfolio exposure analysis.Weather Developer Portal
PerilIQ Storm Data API	Address-level storm history	Hail and other severe-weather event history for U.S. addresses	Useful for lighter-weight integrations and property screening.Peril IQ

My shortlist by insurance use case

1. Automated claim adjudication

Choose:

• Xweather Hail Archive
• Verisk Benchmark

Typical workflow:

1. Receive claim address + date of loss.
2. Query hail events around the property.
3. Compare:

• hail size
• distance from storm track
• event timestamp
• severity threshold

4. Attach weather evidence to the claim file.

Xweather Verisk Integration### 2. Catastrophe modeling / portfolio risk

Choose:

• NOAA MRMS/NEXRAD
• Weather Company HailVision

These are better if you need to calculate:

• hail exposure by ZIP code or portfolio
• accumulation risk
• historical frequency maps
• event footprints

Weather Developer Portal### 3. Quick proof-of-loss screening

Choose:

• PerilIQ
• HalfMileHail

These are easier to integrate but may be better suited for screening than final claim evidence. HalfMileHail, for example, notes that its radar detections are grid-based and should not be treated as proof that hail hit a specific rooftop.Half Mile Hail

Data fields I would require for a claims-grade API

At minimum:

• Latitude/longitude or insured address
• Date/time of hail event
• Hail size (inches/mm)
• Probability/severity score
• Storm footprint or radius
• Distance from insured location
• Source methodology (radar, observation, modeled)
• Historical coverage period
• API availability and audit trail

If you are building a production claims system, I would generally evaluate Xweather + NOAA MRMS together: a commercial API for operational workflows and radar data for deeper validation/modeling.Xweather+1

I can also help compare these by coverage years, API pricing, spatial resolution, and suitability for U.S. property claims if you have a target volume (for example, 10k vs. 1M claims/year).

Yes. For insurance claims analysis, I’d prioritize a source that provides historical, radar-derived hail observations or estimated hail size, rather than a general historical weather API.

Best options

• Tomorrow.io — Their historical API supports point/polygon queries and long historical periods, but their documented hailSize, hailProbability, and hailScore fields are primarily forecast-oriented rather than a long-term historical hail archive.Tomorrow.io Weather API Docs+2
• Xweather — Probably the closest fit to your requirement. Its Hail Archive API provides historical hail events, including hail size and intensity, and is explicitly positioned for applications including insurance. Current documented coverage begins Jan. 1, 2024 in the U.S., with 24-hour query increments.Xweather+1
• The Weather Company / HailVision — A strong option if you need radar-derived spatial hail footprints rather than simply a hail/no-hail value at an address. HailVision provides MESH observations for the contiguous U.S. and Canada at approximately 1-km resolution, including vector data showing areas exceeding specified hail-size thresholds.Weather Developer+1
• NOAA/NWS MRMS — Worth considering if you want to build your own historical hail dataset. The MRMS MESH (Maximum Estimated Size of Hail) product has roughly 1-km spatial resolution and 2-minute temporal resolution. This can provide much richer event footprints, but you'll have substantially more data engineering to do compared with a commercial insurance-oriented API.VLab
• HalfMile Hail — A simple U.S. screening API based on NOAA NEXRAD hail detections. It can return historical hail detections for an address or coordinates, but the provider explicitly warns that its ~0.1° grid is not sufficient to establish that hail actually struck a particular property, so I would not use it as the sole evidentiary source for claim adjudication.Half Mile Hail

What I'd choose for claims analysis

If your workflow is something like:

claim address + date → did a hail event occur nearby? → estimated hail size → event footprint → evidence for claim I'd shortlist Xweather Hail Archive and Weather Company's HailVision first.

The distinction is important:

One caveat: radar-derived hail is evidence of meteorological hail conditions, not proof that a particular roof was damaged. For defensible claims analysis, I'd preserve the raw observation, timestamp, estimated hail size, distance from the property, radar/source metadata, and spatial footprint rather than reducing the result to a simple boolean.

If you tell me (1) U.S. only or global, (2) how many years of history you need, and (3) whether you need address-level results or GIS/event polygons, I can narrow this down to the best 2–3 APIs and compare coverage, pricing, API format, historical depth, and claim-use suitability.