Get a translated PDF that still looks like the original, columns and headings included.
A brochure, a product sheet, a laid-out report: the translation takes the place of the original text, in the same blocks, in the same colours. Where the document's own font cannot write the language you asked for — a French character set rarely holds an « ł » —, a replacement is picked from the real measurements of the original, the same one across the whole document. The file is processed on our servers in Europe, then erased.
This tool needs a machine: your file is sent to our servers in Europe, processed, then automatically erased.
Drop the document, pick the target language, and get back a PDF where the translation sits where the original text was: same blocks, same columns, same colours. The file is sent to our servers, because rebuilding a document means having it.
This tool needs a machine: your file is sent to our servers in Europe, processed, then deleted automatically. We keep no copy and we do not read it.
No, and nobody can promise that: a translated sentence is rarely the same length as the one it replaces — French runs about a fifth longer than English. When text no longer fits, leading tightens first, then the type size is trimmed, and only as a last resort beyond what an eye accepts. What is guaranteed is that nothing overlaps.
A PDF embeds only the characters it uses. A French document holds no Polish « ł » and no Russian « щ », so its own font simply cannot write the translation. A replacement is chosen from the real measurements of the original — serifs, pitch, weight, letter widths — and the same family serves the whole document, so that two neighbouring headings do not end up in two different typefaces.
Those our fonts can write and the layout engine can set: Latin-alphabet languages, Greek and Cyrillic. Chinese, Japanese, Korean and Arabic are not offered — the first three need fonts of several mebibytes, and the last is written right to left, which is a layout problem rather than a font one. For those languages, the text translation tool is still there.
No. The extracted text is sent to the model for the duration of the job, then discarded, and the file is deleted from our servers. Nothing feeds any training.