Word Count
Word Count
For quoting and effort estimation: count segments and characters across several files at once, separate duplicates, non-translatable content and TM match buckets, then export an Excel report.
Steps
- File tools → Word count
- Pick the source language (used for non-translatable detection and TM matching direction)
- Add one or more files in the left panel
- Click Start analysis
- Review the summary and per-file tables; click Export Excel if needed
Supported Files
- Resource files:
.json,.strings,.stringsdict,.xml,.po,.pot,.properties,.csv,.srt,.yaml,.yml - Office documents:
.xlsx,.docx,.pptx - Text and web:
.html,.htm,.md,.markdown,.txt - Bilingual files:
.xlf,.xliff
Counting Categories
Results report segments and characters per category, with their share of the total:
| Category | Meaning |
|---|---|
| Cross-file duplicates | The same source text appearing in more than one of the selected files |
| In-file duplicates | Repeats inside a single file |
| Non-translatable | Pure numbers, symbols, placeholders and similar |
| TM 100% | Identical to an entry in the local memories |
| TM 90%-99% / 80%-89% / 70%-79% / 60%-69% | Fuzzy match buckets against the memories |
| New words | Neither duplicate nor matched — the part that really needs translating |
Results
- Summary: totals across all files, with segments, characters and shares per category
- Per file: one table per file, to spot outliers
- Timestamp: when the analysis ran, useful for reconciliation
Changing files or the source language clears the results; run the analysis again.
Export
Export Excel writes a report with the summary and per-file details to a location you choose — convenient to attach to a quote or delivery note.
Use Cases
- Quoting: price new words and match buckets separately
- Scheduling: estimate real effort from repetition and TM hit rates
- Delivery notes: show clients how much of a file is reusable
Notes
- Local files only: the tool reads files from disk and uploads nothing
- Segmentation matters: resource files segment by key, documents by paragraph, so counts differ slightly by file type
- TM matching is a snapshot: later memory updates do not change an already exported report
- No quality assessment: only volume and match distribution, not translation quality
FAQ
Q: Why is the "new words" share so high?
A: Check the source language, whether a memory covers that language pair, and whether the files belong to the same batch (cross-file duplicates only count when analyzed together).
Q: Are tables inside Word files counted?
A: Yes, table content is extracted as paragraphs and counted; the granularity depends on the document structure.
Q: Can results be saved for later?
A: Export the Excel report — the tool itself keeps no analysis history.
Q: How many files can be analyzed at once?
A: No hard limit, but more files take longer. Batch by project for practical runs.