Skip to content
FertiliCore

Integration platform

Standards in, standards out.

A clinic is never one system. FertiliCore speaks the protocols your hospital, your PACS, your analysers and your registry already use — as five independent connectors on one shared, auditable backbone.

One backbone

Five connectors, one place to look when something goes wrong.

Each connector runs as its own service, so a misbehaving analyser feed cannot take the clinical application down with it. What they share is the part you need on the day an interface breaks: the log, the crosswalk and the queue.

Every message logged

Direction, type, external reference, payload, status, error and correlation id — for inbound and outbound alike.

Identifier crosswalk

A record of how each external identifier maps to an internal patient file, so a foreign record is never silently duplicated.

Reliable outbound queue

Claim-locked, retried on failure, reclaimed if stuck, and visible to operators with a manual retry.

Tenant-scoped throughout

Connector configuration, message log and crosswalk all sit behind the same row-level isolation as clinical data.

Nothing is on by default.

Installing FertiliCore changes nothing about what your network talks to. Every connector arrives disabled. Every endpoint and every credential arrives empty. There is a master switch above all five, and it is off. Until an administrator configures a transport and turns it on, the integration services sit idle and say so in the log.

Inbound listeners bind no socket until configured
Peer and application-entity allow-listing on inbound transports
Optional TLS on inbound listeners
Service-to-service authentication to the clinical API
Secrets encrypted at rest, never returned in plaintext
Per-tenant configuration — one clinic cannot see another's endpoints

Integration questions your IT team will ask

Does anything start talking to our systems the moment FertiliCore is installed?

No. Every connector ships disabled, with no endpoint and no credential configured, behind a master switch that is also off. Listeners bind no socket and watch no folder until an administrator configures and explicitly enables them. Nothing leaves your network by accident.

Where are integration credentials stored?

Connector settings live in the system settings store. Secrets — credentials, tokens, API keys — are encrypted at rest and are never returned in plaintext by the administration API; the interface shows them only as set or unset, and they can be overwritten but not read back.

What happens when the partner system is unavailable?

Outbound messages go through a reliable queue rather than a direct send. Entries are claimed with row-level locking so multiple instances never double-send, failures are retried, a stuck send is reclaimed automatically, and operators can watch and retry the queue from a dedicated screen.

Can we see exactly what was sent and received?

Yes. Every inbound and outbound message is recorded with its direction, type, external reference, full payload, processing status, any error and a correlation identifier. An external-to-internal identifier crosswalk records how each foreign record was matched to a patient file, so a mismatch can be traced rather than guessed at.

Do the integration services write to the clinical database directly?

No, and this is deliberate. Each connector calls the clinical application's own ingestion API rather than writing tables. An externally sourced record therefore passes through the same validation, the same permission checks and the same audit trail as one typed in by a clinician.

Send us your interface specification

Tell us which HIS, PACS, analysers and registry you run. We will come back with a written statement of what maps cleanly, what needs work, and what we cannot do.