Context-rich matches
Show previous usage with domain, document type, language pair, source, audience, and approval status.
Feature
Translation memory
Reuse accepted language while keeping track of domain, register, source, and review confidence.
Reuse with judgment.
Translation memory can save time, but blind reuse can spread old mistakes. Idanium’s memory layer connects repeated phrasing with domain labels, source quality, review status, and audience constraints.
That gives translators signal, not just matches: what was used, where it worked, who approved it, and whether the context still fits.
Memory hygiene
Flag stale, conflicting, or low-confidence segments before they reappear.
Reusable patterns
Promote frequently approved phrasing into templates, glossary notes, or guardrails.
Review confidence
Separate machine suggestions, draft matches, expert-approved segments, and deprecated language.
What this adds
Memory becomes evidence, not only storage.
A match is only useful when the user can judge whether the previous context still applies.
Exact reuse
Approved phrases can be reused quickly when domain, audience, and document type match.
Near match
Similar content can suggest a starting point while highlighting differences that need review.
Conflict check
Older memory can be compared against current glossary rules and source changes.
Connected records
Each memory segment carries operational metadata.
This metadata is what keeps translation memory from becoming a pile of old sentences.
Source context
Document type, original date, source authority, version, and citation.
Review state
Draft, reviewed, approved, deprecated, conflicting, or needs-domain-review.
Audience
Expert, professional, public, learner, patient, resident, or child-oriented language.
Term links
Glossary entries and ontology concepts that appear inside the segment.