email   password     Lost password?

Wordfast Aligner is offered to the translation community as a free service. We designed it with usability in mind: copy-paste text, download a TM. Doesn't get any simpler.

Back to Wordfast Aligner

List of supported languages

We support any two-language combination of:

Arabic, Bosnian, Bulgarian, Catalan, Chinese (simpl. & trad.), Croatian, Czech, Danish, Dutch, English, Estonian, Finnish, French, German, Greek, Hebrew, Hindi, Hungarian, Indonesian, Italian, Japanese, Latvian, Lithuanian, Polish, Portuguese (incl. PT-BR), Norwegian, Romanian, Russian, Slovak, Spanish, Swedish, Turkish, Ukrainian, Vietnamese.

If your language is not listed, we can add it if you supply us with a Translation Memory of more than 100,000 Translation Units.

  • Target audience This service is intended for freelance, individual translators. It is not intended to process corporate data on an industrial scale. Wordfast Aligner can be purchased for corporate use; it is easy to integrate in any workflow because of its simple command-line interface, small footprint (one single Windows EXE file, plus one module per language), versatility, and industrial-strength capacity.
  • Liabilities Every effort is done to provide a high-quality service; however, no guarantee of any sort is offered or implied on the results.
  • Confidentiality The aligned material (all posted source and target material, and the alignment, or the TM, that is returned) are deleted every day. We do not disclose your material in any way. Our reputation is at stake. Note that our aligner learns from the submitted material by drawing associations between words or expressions, for the sole purpose of improving future alignments.
  • Cost It's free -- no catch, no ads. And it's not limited to Wordfast users for the moment, as we do not check for product license credentials.

  • It's fast. The typical alignment takes under a minute. This is why you download the TM right after alignment, we don't email it to you. We do not collect any personal information here, we don't even use cookies.
  • It's precise. Our aligner uses a huge collection of existing translation memories to create alignment of linguistic grade. Most other aligners are statistical aligners, which rely on non-linguistic clues, and on the parallelism of the submitted document pairs (in other words they align already-aligned documents: wow). Our aligner is linguistic - you can shuffle all paragraphs before submission, the result will be just as good.
  • It aligns sentences, because most Translation Memories are made of sentences. This aligner is not meant to align mere list of words, or lists of terms. For that purpose, you need a terminology extractor.

What can I expect?

Aligners cannot produce a perfect alignment all the time. There are occasional mishaps in an aligned TM. In most TMs actually. Alignment is only as good as:

  • the documents it is based on;
  • the software's capacity to properly segment the two documents. Segmenting means separating a paragraph into sentences, because TMs are usually sentence-based. We use linguistic intelligence in the process, with sets of abbreviations, rules, etc., but the segmentation cannot be perfect all the time. We aim for an accuracy of at least 75%, which provides a substantial boost in translation, but usually achieve a better score.

As a result, some sentences may be unmatched or mismatched. Translation Memories should in general be handled carefully: after all, the translator is the only creator of meaning.

Testing vs Using: Academical vs Professional

Wordfast Aligner caters to professionals first - translators that actually translate to pay the rent. We want to offer a no-hassle design with good productivity. As explained above, aligned TMs offer a boost in the translation process, they are not an end in themselves.

The academical world, and some corporate decision-makers, sometimes scrutinize TMs produced by aligners and are quick to notice a degree of inconsistency. The output of all aligners (and/or terminology extractors) looks dirty - but one should remember what aligned TMs really are for. An aligned TM is typically used as a secondary, read-only TM; on that base, whatever the translator chooses to re-use or edit is then recorded in the primary, read-write TM, sometimes called the gold TM. As such, aligned TMs are backup material, discarded after a project. When done properly, pre-project alignment offers great value, for it is generally superior to Machine Translation. With no malice: if your aim is academical perfection, we suggest that you find and use a perfect aligner - ours is productive, not perfect.

I'm not a translator. What is alignment anyway?

Alignment takes two documents in two languages -- called source and target -- as input. It matches sentences in the source document to sentences in the target document. The output is called a Translation Memory (TM). You can think of a TM as a bilingual dictionary, with sentences instead of words. TMs are used in translation tools to boost productivity. TMs do not translate, as Machine Translation pretends to; but they provide clues and propositions that are checked, corrected, and recycled by translators. TMs relieve translators of repetitive tasks, so they can focus on the creative side of their work. Thus, TMs are useful in technical and commercial translation; much less in literary translation.

Help! I cannot import my aligned TM into my existing TM.

A Translation Memory (TM) is just a text (TXT) file, whether it's a Wordfast Classic file, or a TMX file. In the zip you receive after alignment, those are two numbered files, which look like:

123456.ali.txt(a Wordfast Classic TM)
123456.ali.tmx(a TM in a universal format)

Both can be opened as TXT files. Please use a text editor, like Notepad or TextEdit -- not a browser or Word -- to open those files. If your translation tool fails to import one file type, try to import the other file type. There is a file type selector in WF PRO's Import TM > Select file dialog box, as in most translation tools.

If you are not using a Wordfast translation tool, the TMX file is the one you will import.

The usual problem in importing a TM is compatibility between the language codes set up in the translation tool, and those in the TM you are trying to import. It's the old problem of adding apples and pears: the Translation Units (TUs) that are imported must have language codes matching those in the tool's "receiving" TM. With some tools, those language codes must be identical, and that's often the problem. Some tools have flexibility, only taking the language into account (first two characters, like EN, or en, or FR, or fr), not the region, or locale (last two characters, like US, or GB, or FR, or CA, etc.) Some tools expect a full and strict compatibility of language codes, including locale, some only rely on language, without locale. Also note that most tools consider language codes to be case-sensitive: EN would not match en or En.

The Wordfast aligner lets you specify the language codes before alignment.

It is easy to correct a TM. Perform a search-replace action on language codes with a text editor. With a TMX file, use the surrounding quotes to avoid replacing the wrong text (for example, search for "en" and replace with "EN-US", quotes included, to change en into EN-US). With a Wordfast Classic TM, use the surrounding tabulators instead of quotes.

albanian | arabic | armenian | basque | bulgarian | catalan | chinese (simpl) | chinese (trad) | croatian | czech | danish | dutch | english | estonian | german | finnish | french | greek | hebrew | hindi | hungarian | indonesian | italian | japanese | korean | latvian | macedonian | norwegian | polish | portuguese | romanian | russian | serbian | slovak | slovenian | spanish | swedish | thai | tagalog | turkish | ukrainian | vietnamese | my language | (c) 2017, Yves Champollion
Yves Champollion translation localization consultant consulting. en français