Machine Translation Post-Editing Guide
Fix machine output efficiently.
Added October 15, 2026. Written with AI assistance and checked for clarity and placeholders; not reviewed by a lawyer, accountant, clinician or other professional. How these prompts are made
The Open buttons send the prompt text as currently shown below to that site only when you click. Fill in the placeholders first.
Create a post-editing guide for machine-translated text. Language pair: [LANGUAGE_PAIR]. Content type and risk: [CONTENT_RISK]. Desired quality level (gist, publishable, high-stakes): [QUALITY_LEVEL]. Define what to fix at each quality level and what to leave alone. Provide an error checklist for typical machine mistakes: wrong terminology, missing negation, mistranslated numbers and dates, gender and agreement errors, idioms, and inconsistent names. Add a process: read source and target together, fix meaning errors first, then terminology, then fluency, then final read of the target alone. Include time guidelines per 1,000 words as a range to be adjusted from my own measurements. Add a short sample: a machine sentence with errors, the post-edited version, and notes. State that medical, legal or safety content needs a qualified human review. Add a short training exercise: five machine-translated sentences written by you in the language pair with deliberate errors, followed by an answer key, so editors can practise. Include a recommended record of changes for quality tracking and a note on when it is faster to translate from scratch than to post-edit.
Fill in before use
Optional: type your details and the prompt above updates. Nothing is saved or sent.
Replace each bracketed placeholder with your own details, then paste into Any AI or another assistant.
See an example result (fictional data)
Quality level: publishable. Fix order: meaning, terminology, fluency, final read. Machine: "The patient must not take the medicine with food." (target language dropped "not") Post-edited: restore the negation; flag as critical. Checklist: negation, numbers and dates, names, gender/agreement, idioms. Time: 800-1,200 words an hour for light edits (adjust to your measurements). Medical, legal or safety content: have a qualified human review. Sample: "Take two tablets daily" machine gave "twice daily"; post-edit and note the error type: mistranslation. Edit record: error type, severity, time spent. Rule: if more than half the sentences need rewriting, retranslate instead.
Illustrative only, made-up data. Your results will differ and must be reviewed.