Skip to main content

What is Haste Health?

Haste Health is an open-source, self-hosted FHIR R4 clinical data repository, written in Rust and built for AI agents. In plain terms: point Epic, Cerner, HL7v2 feeds, or any other FHIR-compatible system at it, and you get back one normalized FHIR API that your apps, dashboards, and AI agents can all read and write against.

Comes in from
EpicOracle HealthMEDITECHathenahealthHL7v2 feedsAny FHIR R4 API
Haste Health
FHIR R4 clinical data layer
Repository and search
OAuth 2.0 and SMART
Access policies
Audit trail and history
Goes out to
AI agentsClaude, GeminiApps and portalsBackend servicesAnalytics and BIPartner systems

How it works​

A request is a normal FHIR REST call, scoped to your tenant and project:

GET /w/acme-health/production/api/v1/fhir/Patient?name=williams
{
  "resourceType": "Bundle",
  "type": "searchset",
  "total": 1,
  "entry": [{ "resource": { "resourceType": "Patient", "id": "kin5px...", "name": [{ "family": "Williams" }] } }]
}

No proprietary SDK and no separate ingestion pipeline: the same FHIR API is what other FHIR servers, EHR integrations, and MCP-connected AI agents all use to read and write data.

Why teams use it​

  • Self-hosted, Apache-2.0: own your infrastructure and your data, no vendor lock-in
  • MCP-native: agents get structured FHIR access through generated tools, not a scraped UI or a bolted-on integration
  • Fast: sub-100MB footprint, under 10ms writes, runs without oversized infrastructure
  • Standards-based auth: OAuth 2.0 and SMART on FHIR built in, so you're not building an auth layer from scratch
  • Flexible search: Postgres-only search is available; Elasticsearch is optional, not required

Common uses: EHR integration, patient portals, research platforms, care coordination, and giving AI agents structured access to clinical data.

Next​