Machine Translation Post-Editing
Professional quality control and correction of machine-translated content from DeepL, Google Translate, and custom NMT engines, delivered to publication-ready standard.
About this service
Machine translation has improved substantially in recent years, particularly for high-resource language pairs like EN-ES and EN-FR. But current NMT output still fails in predictable ways: terminology inconsistency across a long document, awkward nominal structures that a native speaker would never use, and errors triggered by source text that is ambiguous or idiomatic. Post-editing addresses these without starting from scratch. Light post-editing corrects fluency errors and clear mistranslations while preserving as much of the MT output as possible. This is appropriate for internal documents, draft content, and gisting. Full post-editing brings the output to the same standard as a human translation, correcting all errors in accuracy, fluency, consistency and register. This is appropriate for any content that will be published or shared with clients or partners. We accept output from DeepL, Google Translate, Microsoft Translator, and custom NMT engines including Opus-MT and custom-trained models. We require the source text alongside the MT output.