Skip to main content

Projects and files

Creating a project​

Project creation is a four-step wizard: Project Info → Upload Files → Configure Processing → Processing.

Project creation wizard

Step 1: Project Info​

  • Project Name and an optional Description.
  • Client Name, Client Industry, and Department — used for organisation and as context for AI translation.
  • Source Languages and Target Languages (searchable lists). Selecting the same language on both sides switches to monolingual mode — use variants like en-US → en-GB.
  • Workflow Type — how much of the Translation → Revision → Proofreading pipeline runs automatically (see Workflow stages):
    • Manual (default) — a human linguist drives each stage.
    • Automatic End-to-End — Translation and Revision run automatically; Proofreading stays manual.
    • Auto Translation Only — Translation runs automatically, then hands off to a human at Revision.
  • Style Guides (optional) — attach style guides relevant to your target languages; they steer AI translation conventions.
  • Research Context Description (optional) — free-text context about the content that feeds project research and AI prompts.

Step 2: Upload Files​

Add the files to translate. Word, Excel, PowerPoint, PDF, JSON, YAML, XLIFF, HTML, IDML, SCORM, and more are supported. Very large files trigger the analysis flow described in Large files.

Step 3: Configure Processing​

The platform detects a Processing Method per file, which you can override:

  • Standard Conversion — normal extraction of translatable text into segments.
  • OCR Processing — for scanned PDFs and images: choose the OCR Provider and optionally supply a Custom OCR Prompt and an Output Format.
  • Format-specific options such as Extraction Instructions for unusual file structures.

Step 4: Processing​

Files convert in the background; progress is shown per file. Once done, files are ready for analysis and translation.

File analysis​

Run File Analysis on a project to get a CAT-style breakdown before translating — word counts by match category: perfect matches, source matches, repetitions (internal and cross-file), non-translatable content, leveraged content, and new content. Use Detect Cross-File Repetitions when a batch shares content across files. Past runs are kept under File Analysis History, so you can compare analyses (e.g. before and after attaching a TM).

This is the basis for quoting: only new and fuzzy content costs full effort.

Counting is script-aware: for Chinese, Japanese, Korean, and Southeast Asian languages, the analysis counts characters instead of space-separated words — so CJK volumes are quoted on the metric the industry actually uses.

Project detail page

The project page​

The project detail page is the working hub:

  • Files — per-file status, workflow stage, language pair, and actions (open in editor, re-import, delete).
  • Stats — overall progress, total words, segment counts, quality and effort scores.
  • Workflow Settings — stage automation, auto-approve threshold, quality threshold, and stage assignees (see Workflow stages).
  • Translation Memory / Term base — attach TMs and TBs, set the TM Threshold (%), and toggle Use in AI Translation.
  • Audit History — who did what, when.

Deleting and restoring files​

Deleting a file moves it to the Recycle Bin rather than destroying it; restore from there if a deletion was a mistake.