Ontoserver and Kodjin both show up on the terminology-server shortlist for EMR vendor backends, and they win deals from different kinds of vendors. Ontoserver leans into terminology depth and edition handling; Kodjin leans into multi-tenant operations and developer ergonomics. The choice between them rarely turns on which one is technically better, because both handle the core operations well. It turns on what the EMR vendor wants to spend engineering time on over the next several years. For broader context, see interoperability vendor evaluations.
The FHIR terminology servers for medical software vendors reference guide covers where the terminology server fits in the broader EMR vendor stack.
What Ontoserver Gives an EMR Vendor
Ontoserver is CSIRO's commercial server, used in national-scale deployments across Australia and adopted by EMR vendors that serve research-heavy customers. The strengths are terminology depth: strong SNOMED CT edition handling, mature reference set behavior, broad LOINC and ICD-10 coverage, and a $expand engine that holds up against large value sets.
The EMR vendor that picks Ontoserver gets a server tuned for terminology correctness above all else. The honest cost is that the operational story leans toward customers who want to run the server themselves in a controlled environment. Ontoserver is most at home in deployments where the customer has terminology expertise on staff and treats the terminology server as a first-class operational dependency.
What Kodjin Gives an EMR Vendor
Kodjin Terminology is a commercial terminology product that targets EMR vendors and integrators directly. The strengths are developer ergonomics: clean APIs, multi-tenant features that fit vendors serving several customers from one deployment, and a managed content update story for SNOMED CT and LOINC.
The EMR vendor that picks Kodjin gets a server tuned for vendor operational needs above terminology depth. The trade-off is that some specialty terminology workloads (rare edition variants, deep reference set use, research-grade SNOMED CT manipulation) are handled better by Ontoserver. For the typical EMR product, Kodjin's defaults match the vendor's needs more closely. The self-hosted vs managed FHIR terminology servers for EMR vendors comparison covers the operational trade-offs that shape the broader self-host versus managed decision.
How EMR Vendors Should Pick
The choice comes down to two questions. The first is whether the EMR vendor's customer base demands terminology depth (research, public health, national deployments) or operational simplicity (community hospitals, specialty practices, telemedicine). Terminology-depth customers push toward Ontoserver; operational-simplicity customers push toward Kodjin. The second question is whether the EMR vendor wants to operate one terminology stack across all customers or per-customer stacks; multi-tenant operation is Kodjin's strength.
In practice, EMR vendors with five or more customers running on a shared backend usually pick Kodjin for the multi-tenancy. EMR vendors with a small number of large enterprise customers, each running its own deployment, often pick Ontoserver for the terminology depth.
For vendors that also serve coding-heavy telemedicine workflows, the best terminology tools for telemedicine coding workflows walkthrough covers where each option sits across those workloads. The defensible choice between Ontoserver and Kodjin is the one that matches how the vendor wants to operate, not the one that scores higher on a feature matrix.
EMR vendors that end up happiest with their pick usually run both options through a two-week evaluation against their actual customer mix before procurement. The differences between Ontoserver and Kodjin in production rarely match the differences in the procurement deck, and the only way to know which one fits is to put real load through both.
Vendor teams that ran a head-to-head pilot consistently report that the procurement winner often is not the engineering favorite, and the conversation between the two perspectives is where the right decision usually lands for the actual customer mix.
Sources
- Ontoserver architecture and capabilities - PMC paper, peer-reviewed, foundational
- SNOMED with FHIR by Peter Groves Williams - PDF slides, DevDays, 2024
- Terminology Module guide - HTML spec, HL7, 2024