Machine translation has got significantly better — but it still produces output that is confidently wrong, culturally off, and stylistically flat in Turkish. Our post-editors fix what the machine gets wrong and leave what it gets right.
We post-edit output from DeepL, Google Translate, Azure Translator, Amazon Translate, and custom neural MT systems. We can also set up an MT engine integrated directly into your workflow.
Light MTPE (MTPE-L) corrects clear errors — wrong words, grammatical mistakes, and factual errors — without rewriting fluent but non-ideal output. Full MTPE (MTPE-F) targets publishable human-quality output, which may require significant rewriting. We advise on the right service level for your content type.
MTPE is priced per word on the source volume. Rates are lower than full HT but vary by content type, language pair, and MT quality. We assess a sample before quoting.
Yes — we work with SDL Trados, memoQ, Phrase (Memsource), Smartcat, and most other major translation management systems, including direct API integration for automated workflows.
Yes — all MTPE projects are run with client-specific TMs and glossaries to ensure consistency across batches and to improve MT output quality over time.
Send a sample — we'll assess MT quality, recommend a service level, and quote within a few hours.