Skip to content
Ever
中文

developers

Build on open healthcare infrastructure

A developer-friendly platform to ensure you're never out of solutions. Standard interfaces in, standard interfaces out, and a plug-in model for everything in between.

Read the docs ↗Request source access
fhir · capability statement
$ curl "$MEDOS/fhir/metadata"

{
  "resourceType": "CapabilityStatement",
  "fhirVersion": "4.0.1",
  "rest": [ { "mode": "server", ... } ]
}

How would you like to get started?

learn

Learn the platform

How the hospital system, the exchange and the patient app fit together.

open ↗integrate

Connect a system

Read and write clinical data over FHIR R4, or feed HL7v2 ADT, ORM and ORU messages.

open ↗run

Run it yourself

A single Docker Compose stack for on-premise sites, including air-gapped ones.

open ↗extend

Build a plug-in

Create your own module, or install pre-built ones from the app store.

open ↗

try it

Standard interfaces, no proprietary SDK required

Examples run against your own deployment: set $MEDOS to its address. There is no shared public sandbox yet.

fhir r4 · international patient summary
$ curl "$MEDOS/fhir/Patient/{id}/\$summary" \
    -H "Authorization: Bearer $TOKEN" \
    -H "Accept: application/fhir+json"
hl7v2 · send a raw message
$ curl -X POST "$MEDOS/hl7/raw" \
    -H "Authorization: Bearer $TOKEN" \
    -H "Content-Type: application/json" \
    -d '{"rawMessage":"MSH|^~\\&|LIS|LAB|MEDOS|HOSP|..."}'

the stack

How a request travels

A microservice backend behind an API gateway, a React application in the browser, and a projection layer that serves as the read model. Writes go to the services; screens subscribe to the read model and update in real time.

service bus
NATS
framework
NestJS + Moleculer
primary store
MongoDB
read model
PostgreSQL
file store
IPFS + S3
frontend
React + Vite
Data flow in detail ↗
request flow
Frontend (React)
   │ REST / HTTPS
   ▼
Load balancer or nginx
   ▼
API gateway        JWT validation, routing
   ▼
NATS message bus   request / reply + pub/sub
   ▼
Microservices
   ├─▶ MongoDB           primary write store
   ├─▶ projection        read-model tables
   │     └─▶ Frontend    realtime subscription
   ├─▶ IPFS / S3         files
   └─▶ External systems  FHIR / HL7 / SMART

interfaces in detail

FHIR R4

Read and write operations with subscription-based event delivery. SMART on FHIR apps talk to the platform through the public API service.

read

GET and search for Patient, Encounter, Observation, Condition and more.

write

POST, PUT and PATCH, mapped back into the platform's own records. Off by default per hospital.

subscribe

Register criteria; matching events are delivered as FHIR Bundles, HMAC-signed.

discover

A CapabilityStatement at /fhir/metadata lists exactly what a deployment supports.

HL7v2

For laboratory instruments, admission feeds and pharmacy systems that speak the older standard.

transport
MLLP over TCP, or REST
default port
2575
admissions
ADT A01 to A08
orders
ORM O01
results
ORU R01
replies
ACK / NAK for every message
FHIR guide ↗HL7v2 guide ↗

self-host

One command to a running hospital stack

The on-premise bundle is a single Docker Compose file: database, message bus, file store, every service, a reverse proxy and the frontend, with health checks and startup ordering.

  1. Generates a TLS certificate for your hostname.
  2. Creates environment files with random secrets.
  3. Sets language and addresses for the chosen country.
  4. Builds all images.

Source access is by application while the public release is staged.

install
$ cd infrastructure
$ ./scripts/setup-onpremise.sh     --region japan --hostname medos.local --yes
country packs
market-packs/
├── medos-thailand/      th-TH · Asia/Bangkok · NHSO
├── medos-japan/         ja-JP · Asia/Tokyo  · Kaigo
└── medos-philippines/   fil-PH · Asia/Manila · PhilHealth

Explore the documentation

Documentation is published with the product at evermedos.com.

foundations

interoperability

deploy and localise

app store

Future-proof, ever-changing: the Ever App Store

An app store for plug-ins: create your own and offer it on the store, or install pre-built solutions for urgent needs.

Open-source programme ↗RoadmapCommunity
“Our hospital recently shifted to operating under a value-based care model, so the developer team quickly deployed a new plugin onto MedOS that allowed for the collection and reporting of patient outcomes and patient satisfaction.”
“We downloaded Ever's ChatGPT plugin for document generation to help the administrative team quickly draft out various paperworks. We are currently testing out another AI plugin that allows new team members to self-onboard.”

Questions developers ask

Is there a public sandbox?

Not yet. Examples run against your own deployment. Write to us for evaluation access.

Do I need a proprietary SDK?

No. The external interfaces are FHIR R4 and HL7v2, so any standard client works.

Can I write data, or only read?

Both. FHIR write is switched off by default and enabled per hospital, because writing into a clinical record is that hospital's decision.

How are events delivered?

Through FHIR subscriptions and webhooks. Deliveries are signed so the receiver can verify the sender, and rate-limited per subscription.

Is the source open?

The public release is being staged. Until then source access is by application, with free-use licences for qualifying public-health programmes.

Which languages does the interface support?

Seventeen locales, including Thai, Japanese, Filipino and English.

Join the transformation today.

Ever was created from the ground up in partnership with pioneering engineers and expert physicians. Our innovations solve real-life issues that inhibit medical institutions around the world.

Contact us →