The benefits data API.
The Hillwinds API returns employer benefits data as structured JSON — companies enriched from Form 5500 filings, the broker office and carriers behind each one, the people who make the benefits decision, and the signals that say a change is coming. One authenticated request, no CSV wrangling, no filings to parse.
What you can query
Seven entities, one benefits graph.
Every endpoint resolves against the same joined dataset, so a company query and a broker query return two views of the same relationships rather than two disconnected tables.
/v1/companiesEmployers enriched from Form 5500 filings — plan year, premiums, funding type, employee bands, industry, and the broker and carrier relationships attached to each.
/v1/personnelValidated HR, Total Rewards, Finance, and leadership contacts mapped to their employer, filterable by seniority, function, state, and email availability.
/v1/brokersBroker firms with their book of business — which employers they serve, at what size, and how that book has moved across plan years.
/v1/broker-officesOffice-level rollups beneath each broker firm, so a national brokerage resolves to the branch that actually holds the relationship.
/v1/lookalikesGiven a set of companies, return the accounts that most resemble them across the benefits graph rather than generic firmographics.
/v1/companies?signals=Predictive change signals — broker-of-record changes, carrier switches, moves to self-funding, and PEPM swings, flagged ahead of the renewal.
/v1/filter-optionsThe controlled vocabularies behind every filter, so you can build valid queries without guessing at enum values.
Example request
A first call, start to finish.
Bearer token in the header, filters in the query string, JSON back. Test keys read sandbox data and charge zero credits.
curl -H "Authorization: Bearer $HILLWINDS_API_KEY" \ "https://api.hillwinds.ai/v1/companies?states=NY&employee_bands=Mid-market+Accounts&page_size=3"
{
"ok": true,
"data": [
{
"id": "737336823",
"company_name": "Copperleaf Financial Co",
"company_state": "NY",
"employee_band": "Mid-market Accounts",
"total_number_of_employees": 481
}
],
"meta": { "credits_charged": 0 }
}Three ways in
REST, MCP, or bulk.
The same graph, shaped for whichever way your stack consumes data.
Bearer-token authenticated JSON over HTTPS. Cursor pagination, field tiers, and per-request credit accounting in the response headers.
Read the API reference →The same graph as a Model Context Protocol server, so Claude and ChatGPT can query benefits data directly in a conversation.
AI agent guide →Bulk delivery for warehouse loads and enrichment jobs that would otherwise page through millions of records.
Flat-file mode →API FAQ
Questions developers ask first
What data does the Hillwinds API return?+
Employer records built from Form 5500 filings and enriched with broker offices, carrier relationships, funding type, premiums, validated benefits personnel, and predictive change signals. Every record is queryable by 40+ benefits-native filters rather than generic firmographics.
Is there a Form 5500 API?+
Yes. The Hillwinds API is a commercial Form 5500 data API. It serves normalized filing data joined to company identity, broker and carrier relationships, and benefits personnel — rather than the raw DOL EFAST datasets, which arrive as annual dumps you parse yourself.
How is authentication handled?+
A bearer token in the Authorization header. Every account gets a live key and a test key. Test-key reads run against sandbox data and charge zero credits, so you can validate an integration before it costs anything.
How does pricing work for API access?+
Requests draw down a credit balance, and the cost of each request is returned in the response headers so usage is auditable per call. Free operations — filter options, autocomplete, and sandbox reads — do not consume credits.
Can AI agents query Hillwinds directly?+
Yes. Beyond the REST API there is an MCP server, an endpoint index and recipe library written for LLM tool use, and a full documentation corpus published at llms-full.txt for retrieval.
Keep reading
Related documentation
Form 5500 API
The commercial API over normalized DOL Form 5500 filings, and how it differs from raw EFAST datasets.
AI agent guide
Endpoint index, autocomplete reference, and a recipe library written for LLM tool use.
Filtering & search
The 40+ benefits-native filters, how named filters differ from column filters, and when to use each.
Credits & usage
What each request costs, which operations are free, and how to read the credit headers.
What is Form 5500?
Plain-English guide to the annual ERISA filing that the company graph is built from.
Data & signals
Where the underlying data comes from, how it is verified, and what the predictive signals detect.
Ready to talk?
Build on the benefits graph.
Get an API key, or see the data on your own book of business in a 25-minute demo.