Importing your first PDF
Documents is where a file becomes a translation you can read at the reading desk or download. This guide takes one file from the drop zone to the finished copy, and says which choices actually matter along the way.
What you start with
Open Documents from the sidebar. The Import tab accepts PDF, DOCX, EPUB, TXT, MD and MOBI files, and the drop zone lists the size and page limits that apply. Image files such as JPG or PNG are not accepted here: a scanned document has to arrive as a PDF.
Nothing is uploaded when you pick a file. The browser first checks the format, counts the pages and estimates how much of a PDF already has a text layer, so a file that is too long or not supported is turned away before it costs anything.
Step by step
- 01Drop the file inDrag it onto “Drop a file here”, or press “Choose a file”. A bar appears with the file name, the page count and a badge: “Ready to translate”, or “Needs OCR” when the browser thinks some pages are scanned.
- 02Choose the languageUnder “Translate into”, pick the language you want. You never pick the source language: it is worked out automatically as the document is translated.
- 03Choose the output version“Output version” has two cards, “Bilingual” and “Translation only”, and one of them always stays on. For a PDF, “Bilingual” brings back two files from the same run, the bilingual PDF and a translation-only PDF, while “Translation only” on its own brings back just the translation. Word, EPUB, TXT and Markdown files come back as one document in their own format, bilingual or translated alone.
- 04Check the rows under “Before importing”The “Apply glossary” row attaches one of your term lists, and “Switch ▾” picks another. “Keep the original layout” is always on: tables, formulas and captions go back where they were.
- 05StartPress “Start translating”. The view moves to “Tasks and delivery”, where a progress bar runs and the upload can be cancelled. If your free pages and credits do not cover the job, you are told the cost before anything starts.
Things to know
Switching “Translate” off changes the button to “Import to reading desk”. The file keeps its original language and no model is called, which suits a document you only want to read and annotate.
If a job fails, the credits it reserved are returned. “Failure report” on the task explains what happened, and “Retry” runs it again when that is possible.
Uploaded files and their translated copies are kept for 7 days and then deleted from storage, so download anything you want to keep.
What comes next
A finished task is filed in My assets, and downloading it there lists the files it has: “Bilingual PDF” and “Translation only” for a PDF, depending on the version you chose, or “Translated document” for a Word, EPUB, TXT or Markdown file. A MOBI ebook is not written back as a file; it is read at the reading desk only. “Open in reader” takes you to the desk with the translation and the source side by side, and “Download” there offers the same files. At the desk you can mark passages, ask the AI about them, and save terms for the next document.
Questions people ask
Do I have to pick the source language?
No. You only choose the language to translate into; the source language is worked out automatically.
Can I get both the bilingual and the translation-only copy?
For a PDF, yes: keep “Bilingual” on and the same run produces both files. With only “Translation only” on, you get the translation alone. Word, EPUB, TXT and Markdown files come back as one document, and a MOBI ebook is read at the reading desk.
How long are my files kept?
7 days. After that the upload and its translated copies are deleted from storage, so download what you want to keep before then.