Skip to content
LionrapidLionrapid Docs

Supported file formats

A standard project starts from one or more source documents. This page lists the formats the upload step accepts, the limits that apply to every file, and what the platform tells you when a file cannot be used. The rest of the chapter explains what happens to each format after upload:

Formats you can upload

Section titled “Formats you can upload”
Kind Extensions Notes
Word-processor documents .docx, .odt, .fodt, .rtf Delivered back in the same format
Plain text .txt One row per paragraph or sentence
PDF .pdf Text-based PDFs give the best result; see PDF documents
Spreadsheets .xlsx, .ods, .fods, .csv Visible cells are translated; see spreadsheets
Presentations .pptx, .ppt, .odp, .fodp Slides, notes, layouts and masters
XLIFF .xlf, .xliff As a source file, or as a pre-translation attached to another file

Every file is delivered back in the format it was uploaded in, with the translation in place of the source text. You never receive a converted copy in a different format.

A few common types are not accepted as project files:

  • Legacy Excel (.xls) — save the workbook as .xlsx, .ods or .csv first. The upload step says so explicitly.
  • HTML, XML and DITA — these are being prepared and the upload step reports them as “coming soon”. Until then, translate web copy through continuous localization.
  • Scans and images (PNG, JPG, TIFF) are the next formats to arrive, with OCR that reconstructs a scanned page into editable rows. Until then the platform works from files with selectable text; see PDF documents.
  • Subtitle files (.srt) are not project documents; they belong to Subtitle Studio.
  • Software string files (JSON, YAML, gettext, resource files) are not uploaded as documents either. Their content flows through a continuous project, where JSON upload is supported in the hub.

Formats not in either list — Markdown, InDesign, Apple iWork, legacy .doc — are rejected as unsupported. Export or save them to one of the formats above.

Limits

Section titled “Limits”

Two limits apply to every source file, whatever its format:

Limit Value
File size 50 MB per file
Length 100,000 words per file

There is no page limit. The size check happens at upload; the word count check happens when the file is analysed, so a very long document is accepted by the upload box and then fails processing with a message naming its word count. Split such a document and upload the parts.

Word counts use the same rule everywhere on the platform, including billing; see How words are counted.

Why a file was rejected

Section titled “Why a file was rejected”

The upload step checks each file before anything is stored. These are the messages you can see, and what to do about them:

Message Cause What to do
Unsupported file type: .ext The extension is not in the list above Save the document in a supported format
Spreadsheet uploads currently support .xlsx, .ods, .fods, and .csv A legacy .xls workbook Save as .xlsx
.ext support is coming soon HTML, XML or DITA Not available yet
‘name’ is actually a PNG image, not a .docx file The file’s contents do not match its extension Upload the original document, or rename the file to its real type
‘name’ contains executable program code The file is a program, not a document Not accepted
Scanned documents aren’t supported yet The file has no selectable text layer Use a digitally created PDF or an editable format
File too large (n MB). Maximum allowed size is 50 MB. Over the size limit Compress images in the document, or split it
File exceeds word count limit Over 100,000 words Split the document

The content check only rejects clear mismatches, such as an image or a spreadsheet renamed to .docx. Text files with a slightly wrong extension, such as a tab-separated file named .csv, are accepted and handled by their contents.

After upload

Section titled “After upload”

Each file is processed in the background before it becomes editable. The file card in the wizard and the hub shows the current stage:

  1. Converting format — the document is opened and normalised.
  2. Parsing document — text, structure and formatting are extracted into rows for the editor.

For files whose source language you asked the platform to detect, a Detecting language stage follows. When processing finishes the file’s word count appears and its targets can be filled by machine translation.

If processing fails, the file card shows the error and a Retry button. Retry after fixing the file, or when the failure looks transient. Errors from inside the document itself, for example a corrupt package or a workbook that cannot be opened, are reported as they occurred, so the text can be technical. The most common fix is to open the file in its native application, save it again, and re-upload.

Section titled “Related”