Machine translation, human judgment
MT engines are fast and getting better every year. But they still confuse registers, hallucinate terminology, and miss context that any domain expert would catch. Our post-editors know where the machines fail — and fix it before your audience notices.
Light PE vs. Full PE
Raw MT
Unedited machine output. Grammatically passable, but full of false cognates, wrong register, and invented terminology. Not suitable for publication.
Light post-editing
Corrected for factual errors, mistranslations, and critical terminology. The text is accurate and understandable, though style may vary from human norms.
Full post-editing
Edited to human-quality standards. Terminology, register, style, and flow all match what a professional translator would produce. Indistinguishable from human translation.
Which tier fits your content
What post-editors catch
False cognates
Words that look similar across languages but mean different things. MT picks the wrong one more often than you'd expect.
Wrong register
MT defaults to neutral tone. A legal brief needs formal language; a marketing email needs conversational. Both get flattened.
Terminology drift
The same term translated three different ways in one document. MT has no memory between segments unless you force it.
Cultural mismatch
Idioms translated literally, humor that falls flat, references that make no sense in the target culture.
ISO 18587 Certified
Our post-editing process follows ISO 18587:2017 — covering post-editor qualifications, defined quality tiers, and documented evaluation criteria. The standard exists because 'just clean up the MT output' is not a process.