Glossary Extractor

Users upload JSON, XLIFF, Excel, CSV, or plain text files and receive a structured term base rather than a simple word list. Each entry includes a clear definition, part of speech where relevant, an exact occurrence count, and translations when the source files already contain them. The resulting CSV file is formatted for direct use in translation workflows and CAT tools. Accuracy comes from separating a fast initial scan that surfaces potential terminology from a subsequent expert review that removes noise and supplies context-aware definitions. This approach avoids over-reliance on any single model while keeping counts precise and flagging non-translatable items such as brand names. The service supports projects of varying sizes, lets users adjust the number of terms extracted, and provides an interactive table for final edits before download. Even single-language outputs retain the rich metadata that distinguishes a term base from a basic glossary.