Most files upload to wxrks and translate correctly with the default settings. But when a file's structure is unusual — a CSV that uses commas instead of semicolons, a JSON export with embedded translator notes, a DOCX with content you don't want translated — wxrks lets you fine-tune exactly how it reads the file and rebuilds it after translation.
​

💡 Who is this for? This guide is for Account Admins and Project Managers who need to fine-tune how wxrks parses and extracts content from a project's files within the wxrks platform.

These options live inside the Upload Files step of project creation. For the full step-by-step of creating a project — organizational unit, languages, workflows, and everything else — see How to create a translation project in wxrks. This article covers only the file-parsing options you'll find in that step.
​

Once a file finishes uploading, click its row to expand a settings panel tailored to that file's type. Click the row again to collapse it.

A few controls appear at the top of every file's settings panel, regardless of file type:

The toolbar above the parameter list — the save icon, the Filters dropdown, and the apply-to-all and clear icons — lets you save the current settings as a reusable filter, apply one you've already saved, push a configuration to every file of that type in the project, or reset a file back to defaults. Reusable filters are covered in How to customize file parameters — this article focuses on what each individual parameter does.

wxrks tailors the parameter panel to each file's type. Complex formats like DOCX, PPTX, and XLSX can expose 30 or more individual fields — every field has an info icon with a full description, so the sections below focus on the ones you're most likely to actually need, grouped by file family.

These formats share several common controls: Aggressive Tag Cleanup (strips extra formatting tags, most useful for documents converted from another format like PDF) and Translate Document Comments (exposes the document's own comments for translation), plus toggles for document properties, headers/footers, hidden content, external hyperlinks, and replacing line breaks or tabs with literal characters. DOCX and XLSX also let you include or exclude text by font color or highlight color.
​

Each format adds its own options on top of that: PPTX adds toggles for master slides, speaker notes, and slide numbers; XLSX adds hidden rows/columns, diagram data, per-worksheet configuration, and treating a file as multilingual; VSDX (Visio) adds toggles for translating specific Visio XML tag types.

HTML exposes two escaping toggles that control how greater-than characters and double-quotes are written back into the translated file — turn off Escape double-quotes if your target system already handles quote escaping and you're seeing characters encoded twice.
​

XML (and DITA, which uses the same parser) controls character escaping, CDATA handling, and whitespace preservation — useful when the receiving system is strict about exact XML formatting.
​

JSON gives you control over which key/value pairs get extracted: Extract all key/strings pairs is on by default, with a regular expression field to exclude specific keys; Infer tags from regex and the two escaping toggles (forward slashes, whitespace characters) control how inline formatting is preserved.

Some JSON exports already carry embedded notes for translators — for example, a key literally named comments that a developer or CMS added next to a string to explain where it appears or how it should read. Instead of leaving that context buried in the raw file, wxrks can surface it directly to the translator and optionally reuse it to improve AI/CS translation suggestions.
​

wxrks looks for a comments key anywhere in the file — at the top level, nested inside an object, or repeated inside a list. This is controlled by two independent flags in the file configuration panel:

You can turn either flag on independently — enabling one does not require the other. Both are off by default.

For example, a source file like this:

With Show Comments In The Editor turned on, the comment text appears next to the "Hello, {name}!" segment: click the info icon on that segment to open a small panel listing the comment.
​

If a comment also contains a link to an image, wxrks automatically downloads it and attaches it to the segment as a reference image, the same way reference images from other file formats appear in the editor — no extra configuration needed beyond having Show Comments In The Editor turned on. That image link must be a public https:// URL pointing to a PNG or JPEG under 5 MB; links that resolve to a private or internal network address are rejected for security reasons, and up to 5 reference images are imported per file.

You can also save these two flags — along with the rest of a JSON configuration — as a reusable filter under your Organizational Unit, instead of setting them on every upload. See How to customize file parameters.

When you upload an already-translated or partially-translated XLIFF file (from another CAT tool, for instance), three toggles control what wxrks does with the existing target content: Import target segments brings the target text already in the file into wxrks, Confirm imported target segments automatically marks that imported text as confirmed, and Save confirmed target segments in TM stores those confirmed segments into your Translation Memory. The remaining options in this panel control lower-level XLIFF writer behavior (CDATA handling, alternative translations, and similar) and rarely need to change from their defaults.

For CSV files, the default field delimiter is a semicolon (;) rather than a comma — change Field Delimiter if your file actually uses commas or another character. Text Qualifier (default: double-quote) lets a delimiter character appear inside a field's value without splitting it. Further down, you can specify which line holds column headers, which columns are IDs versus source/target/comment text, and which columns to exclude entirely.
​

TXT files use the Import As selector itself to pick the parsing mode — plain text by line or paragraph, tab- or comma-delimited table data, or a custom regular expression. Java Properties files (.properties) let you filter which keys get extracted using a regular expression, and control how comments and escaped characters are handled on the way out.

Several less-common formats get their own tailored parameter panel the same way, accessed by expanding the file's row after upload:

For the current full list of extensions wxrks accepts, see Supported and Unsupported File Extensions.

If you run into a configuration this article doesn't cover, reach out to the wxrks support team — you can also search the wxrks Community for prior questions from other users.