Appearance
Skip to content
The right foundation matters.
End-to-end Medplum engineering —
Medplum is the clinical data core —
The next trillion users won't click.
If you're building on healthcare data,
Ready to build agent-native on Medplum?
Medplum Expertise · FHIR-Native · Agent-Native
Build compliant health apps on Medplum — without the guesswork— now agent-native by default.
We help healthcare builders go from concept to production on Medplum's open-source FHIR platform. Implementation, data migration, compliance, interoperability — done right. Plus a native MCP server, a fully spec-compliant FHIR R4 REST API, and a TypeScript Bots framework — clinical data with read AND write surfaces AI agents can act on, today.
100%
Open Source Stack
FHIR R4
Spec Compliant
HIPAA
Compliant Architecture
ONC
Certification Ready
MCP
Native FHIR Server
Why Medplum
The right foundation matters.
Medplum gets the fundamentals right.
Most FHIR servers bolt compliance on as an afterthought. Medplum was architected from the ground up as an open-source, developer-first clinical data repository — and it shows.
Truly Open Source
No dual-licensing traps, no source-available bait-and-switch. Medplum ships its entire platform — under Apache 2.0. You own your infrastructure, your data, and your deployment topology. Enterprise tiers add managed integrations and support, not gated features.
TypeScript-Native SDK
A first-class TypeScript SDK and React component library means your frontend and backend teams speak the same language. FHIR resource types are baked into the type system — catch errors at compile time, not in production.
Compliance by Default
SOC 2, HIPAA, and ONC-aligned out of the box. Access policies, audit logging, and fine-grained RBAC are built into the platform core — not bolted on via third-party plugins.
Interop-Ready Architecture
SMART on FHIR app launch, SCIM, OAuth 2.0, and a Bots framework for event-driven automation. Medplum treats interoperability as a first-class feature, not a roadmap item.
PostgreSQL-Backed FHIR Storage
Unlike FHIR servers that treat JSON documents as opaque blobs, Medplum uses PostgreSQL's JSONB columnar storage to index, query, and filter FHIR resources at the database level. You get the flexibility of document storage with the query performance and reliability of a battle-tested relational engine — no Elasticsearch cluster to babysit.
-- FHIR resources stored as JSONB-- with indexed columns for fast lookupsSELECT id, content FROM patient WHERE content->> 'birthDate' < '1990-01-01'AND compartments @>ARRAY['Organization/acme']; -- No Elasticsearch needed.-- Just Postgres doing Postgres things.
What We Do
End-to-end Medplum engineering —
from first resource to production traffic
We've done this enough times to know where the edge cases hide. Our team covers the full surface area of building on Medplum.
01
Medplum Implementation & Customization
Fresh deployments, FHIR resource modeling, custom Bots, access policies, and project configuration tailored to your clinical workflows.
setup · config · bots02
SMART on FHIR App Development
Build SMART-compliant apps that plug into any EHR ecosystem. We handle app launch flows, scopes, token management, and data access patterns.
smart · oauth · launch03
Clinical Data Migration
Moving from legacy systems, CSV dumps, or HL7v2 feeds? We build ETL pipelines that map your existing data to conformant FHIR R4 resources — clean, validated, complete.
etl · hl7 · mapping04
Healthcare Compliance
HIPAA security controls, ONC certification readiness, audit trail implementation, and policy configuration. We make compliance an engineering discipline, not a checkbox exercise.
hipaa · onc · audit05
Interoperability Services
HL7 FHIR, HL7v2, CCDA, DICOM, and ADT integrations. Connect Medplum to EHRs, labs, pharmacy systems, imaging, and health information exchanges.
fhir · hl7v2 · dicom06
Health Data Stack Integration
Medplum doesn't live in isolation. We connect it to your analytics layer, clinical data warehouse, AI pipelines, and operational ERP — a complete digital health backbone.
tuva · analytics · aiFossible Works Biograph Stack
Medplum is the clinical data core —
but the stack goes further
We integrate Medplum into a full-spectrum digital health platform alongside Frappe for practice management and Tuva for clinical analytics. One partner, one coherent architecture.
Patient EngagementCRM, omni-channel support, website, chatbotfrappe crm
Practice Management & ERPPMS, LIS, RIS, inventory, billing — low-code rapid devfrappe / erpnext
FHIR Clinical Data RepositoryStandardized FHIR R4 storage, SMART apps, compliance, interopmedplum
Clinical Data Warehouse & AIAnalytics DB, dashboards, decision support, notebookstuva / redshift
Agent-Native FHIR
The next trillion users won't click.
They'll call.
Built for AI agents and humans alike.
Medplum was already developer-first. The MCP server, the FHIR REST API, and the Bots automation framework make it agent-first too — clinical data with read AND write surfaces designed for machines.
Medplum MCP Server
A standardized Model Context Protocol surface over the FHIR CDR. Any MCP-aware agent — Claude, ChatGPT, custom — can query, write, and subscribe to clinical resources without bespoke integration code.
FHIR R4 REST API
Every resource type is a versioned, indexed, queryable REST endpoint. Search params, _include, _revinclude, conditional reads, paged bundles — the full spec, exposed as the data plane an agent can rely on.
Bots Framework
TypeScript event-driven automation. Agents hand off long-running clinical workflows to Bots that run server-side with full audit trails, retries, and policy-checked side-effects.
Policy-checked, by default
Access policies, OAuth scopes, and audit logging are evaluated at the API surface — so agent calls are governed by the same RBAC humans are. Safe to give a model write access.
An MCP tool-call. A real FHIR Bundle. Today.
An agent issues a single MCP tool-call. The Medplum MCP server translates it to a FHIR search, runs it through access policies, and returns a spec-compliant Bundle the agent can reason over.
// MCP tool-call from an agent { "tool": "medplum.search", "args": { "resourceType": "Patient", "birthDate": "lt1990-01-01" } } // → FHIR R4 Bundle response { "resourceType": "Bundle", "type": "searchset", "entry": [ { "resource": { "resourceType": "Patient", "id": "pat-019", "birthDate": "1987-03-12" } } ] }
Built For
If you're building on healthcare data,
we can help
Digital Health Startups
Get to market faster with a FHIR-native architecture from day one. Skip months of plumbing and start building features that matter to your users and investors.
Care Delivery Platforms
Telehealth, chronic care management, remote patient monitoring — built on interoperable data that flows between providers, payers, and patients seamlessly.
Patient-Facing Apps
Patient portals, medication trackers, care plan dashboards — powered by SMART on FHIR and built on a data layer patients can actually port between providers.
Clinical Data Pipelines
Normalize messy clinical data into FHIR R4 for analytics, quality measures, research cohorts, and AI/ML model training — with proper provenance and consent tracking.
Let's Build · Agent-Native
Ready to build agent-native on Medplum?
We'll get you to production.
Whether you're starting fresh or migrating an existing system — we'll wire up MCP, the FHIR R4 REST API, and Bots, architect the solution, and ship it with you.