Codon

Platform capabilities

See how Codon handles terminology operations, clinical NLP, governance, and interoperability.

Terminology operations

  • Code lookup with displays, designations and properties
  • Code validation
  • Fuzzy and semantic search with autocomplete
  • Subsumption and transitive-closure queries
  • Batch operations

Cross-system translation

  • Translate codes between any loaded systems
  • Multi-hop translation through intermediary systems
  • More than 1.2 million official built-in mappings across 16 map sets, with additional unofficial derived crosswalks (e.g. UMLS-derived candidate maps) available on request as opt-in, tenant-private content
  • Mapping gap analysis

Code systems and content

  • 28 preloaded Nordic and international code systems
  • Your own custom code systems
  • Supplements: add designations and properties to existing systems
  • Versioned content with full history

Value sets and mappings

  • ValueSet authoring and expansion
  • Mapping workspace with review queues
  • ECL 2.0 queries
  • ConceptMap diff

Clinical NLP

  • Clinical text coding to SNOMED CT and ICD-10 with assertion detection (negation, temporality)
  • De-identification assist that detects and pseudonymises across 56 identifier types
  • Encounter validation against Danish LPR registry rules

AI assists

  • Natural language to ECL, and ECL explanation
  • AI mapping suggestions with justifications
  • ValueSet scope notes
  • Retirement-impact analysis
  • Governance and release summaries
  • Synonym suggestions and change-request triage

Interoperability and conversion

  • HL7v2 to and from FHIR
  • CDA to and from FHIR
  • International Patient Summary (IPS) generation
  • FHIR Subscriptions
  • SMART on FHIR app support

FHIR R4 resources

  • 17 resource types served, including CodeSystem, ValueSet, ConceptMap, StructureDefinition, Questionnaire, Library and PlanDefinition
  • CapabilityStatement and TerminologyCapabilities conformance
  • Profile (StructureDefinition) validation

Import and export

  • 9 export formats: Concepts CSV, Mappings CSV, ClaML, RF2, Epic, HL7v2, SKS, OMOP and Define-XML
  • Bulk data export
  • Delta import

Governance and quality

  • Configurable editorial workflows
  • Release management
  • Concept authoring and retirement
  • Integrity validation
  • Duplicate, overlap and retirement-impact analysis

Terminology insight & the wheel

  • Terminology wheel — a visual map of every code system and the mappings between them, with seven views (structure, coverage, fidelity, licensing, language, freshness, gaps)
  • Per-system quality scorecard — five independent signals (mapping, hierarchy integrity, synonyms, translation, currency), shown separately rather than merged into one number
  • Structural profile — hierarchy depth, branching, leaves, hubs and poly-hierarchy share
  • Mapping coverage and fidelity analysis — how completely and how exactly each map covers its source
  • Detail-loss detection — where many source codes collapse onto one target
  • Consistency checks — flags mapping triangles whose direct and composed paths disagree
  • Coverage asymmetry and one-way-mapping detection
  • Semantic gap detection — suggests related code systems that are not yet mapped
  • Reachability and shortest-path between systems, directly or through chains
  • Language coverage per system (Danish, Norwegian, English)
  • Corpus analytics — coverage and reachability matrices

Security and data protection

  • Per-organisation encryption with cryptographic erasure (GDPR Article 17)
  • Tamper-evident, hash-chained audit log
  • EU AI Act Article 50 transparency records
  • OIDC/SSO, role-based access, two-factor authentication
  • Sovereign mode with no external egress
  • Hosted in Scandinavia

Platform and operations

  • Per-organisation isolation
  • Usage metering and analytics
  • API versioning
  • On-premise deployment option
  • Configurable AI provider, local or your own endpoint

Capabilities vary by plan and add-on — see the pricing page for what's included at each tier. AI suggestions are reviewed by a person before use. AI can run fully locally in a sovereign deployment; the hosted evaluation uses an EU-resident model provider, disclosed in-product.

See plans and what's included→