Arabic machine translation post-editing (MTPE)
Light or full post-editing of Arabic machine translation by native post-editors, under ISO 18587:2017.
Native Arabic post-editors bring machine translation to an agreed level, light or full, under the ISO standard for post-editing.
Arabic machine translation post-editing (MTPE) under ISO 18587:2017, by native Arabic post-editors, for LSPs and enterprises. ISO 18587 is the international standard for post-editing machine translation output: it sets requirements for the full post-editing process and for post-editor competences. El Turco is certified to it and agrees the post-editing level, light or full, before production starts.
Arabic MTPE suits help content, internal documentation and long-tail catalog text, usually in Modern Standard Arabic. Marketing, legal text and UI strings are better served by human translation with a second reviewer, because engines still make gender, number and dialect errors in Arabic. LSPs send MTPE jobs through their CAT tool or portal; enterprises can start with a free pilot on their own content.
Shared workflow: Translation & localization.
Buying for an LSP or an enterprise team? See For LSPs or For enterprises.
Scope
What does Arabic MTPE cover?
- Full post-editing
- The machine draft is brought to the standard of a human translation in terminology, grammar, register and flow. This is the level to agree for anything customers read.
- Light post-editing
- Errors that change meaning or make the text unusable are fixed, and style is left mostly as the engine produced it. Suits internal content and material read once.
- Engine output in your workflow
- For LSP jobs, post-editors work on the MT output in the LSP CAT tool or portal. For direct clients, the engine is chosen for the language pair and content. The client glossary and translation memory are applied before post-editing.
- Sample before scale
- Native linguists post-edit a sample of your own content first and record the error types they find, so the level and the plan rest on your material, not on a generic benchmark.
- LQA sampling
- An LSP or enterprise can add LQA on a sample of delivered segments, scored by category and severity.
Arabic specifics
Where does machine translation go wrong in Arabic?
- 01
Gender agreement
Arabic verbs, adjectives and pronouns change for a male or female reader. When the English source does not say who is addressed, engines pick one and often switch within the same text.
- 02
Dual and plural forms
Arabic has a dual as well as a plural, and the noun form after a number depends on the number. Engines often drop the dual and get plurals after numbers wrong.
- 03
Register and dialect
Engines trained mostly on Modern Standard Arabic produce text that is too formal for consumer content, and they handle dialect input poorly. The brief names the variant, and post-editors hold the text to it.
- 04
Mixed-direction lines
Latin brand names, numbers and variables inside Arabic text can break the line order. Post-editors fix the order in the segment and flag what needs a fix in the layout.
- 05
Digits and punctuation
Engines mix Western and Eastern Arabic digits and leave Latin commas and question marks where Arabic uses ، and ؟. Post-editors apply the digit style and date format set for the locale.
- 06
Terminology drift
An engine can translate the same product term several ways in one file. The client glossary decides, and What's Wrong? checks terms against it before delivery.
Process
How an Arabic MTPE project runs
Turnaround is agreed at scoping, with a 24-hour option for small urgent updates.
Scoping and sample
We review content types, target markets and volume, then post-edit a sample of your own content to see how much correction the engine output needs.
Level and brief
The post-editing level, target variant, digit style, glossary and acceptance criteria are agreed in writing before production starts.
Post-editing
Native Arabic post-editors work under ISO 18587 in your CAT tool or ours, such as RWS Trados, memoQ or Phrase.
QA and sign-off
Like every other delivery, each MTPE job passes What's Wrong? (the translation QA tool from Still Needs a Human) and a human sign-off.
Engine feedback
Recurring engine errors are reported to whoever maintains the engine, so they can be fixed at the source.
Who buys Arabic machine translation post-editing?
Other combinations
Language page
Arabic translation and localization
Modern Standard Arabic and regional dialects for software, games, e-commerce and finance, with RTL testing and Arabic AI training data.
Guide
Arabic dialects and right-to-left layout: a localization primer for product teams
How product managers and designers choose between MSA and Arabic dialects and prepare a mirrored right-to-left interface for launch.
Guide
MTPE vs human translation for Turkish and Arabic: when each works
A decision guide for enterprises and LSPs choosing between machine translation post-editing and human translation into Turkish and Arabic.
Same language
Arabic game localization
Arabic in-game text, right-to-left UI, store pages and in-build LQA for mobile, PC and console titles, in MSA or regional dialect.
Same language
Arabic e-commerce localization
Arabic product catalogs, right-to-left storefronts, checkout and campaigns for Gulf, Egyptian and wider Arabic markets, with MTPE for volume.
Same language
Arabic software localization
Arabic UI strings, right-to-left layout review, help content and release notes, delivered through continuous localization for SaaS and enterprise software.
Same language
Arabic AI training data
Arabic speech collection, transcription, text annotation, RLHF and MT corpora across MSA and the major dialects, labeled by native speakers.
Same language
Arabic linguistic quality assessment
Native Arabic reviewers score translations from any vendor by error category and severity, in Modern Standard Arabic or a named dialect.
Same service
Turkish machine translation post-editing
Light or full post-editing of Turkish machine translation by native post-editors, under ISO 18587:2017.
FAQ
Arabic machine translation post-editing FAQ
What is Arabic MTPE?
Arabic MTPE, or machine translation post-editing, means a native Arabic linguist edits machine translation output until it meets an agreed quality level. El Turco runs Arabic MTPE under ISO 18587:2017, the standard for post-editing machine translation output. Light post-editing fixes errors of meaning only; full post-editing brings the text to the standard of a human translation.
Which Arabic content suits MTPE?
Arabic MTPE suits help content, internal documentation and long-tail catalog text once the engine has been tested on a sample of your content. Arabic marketing, legal text and UI strings stay in human translation with a second reviewer, because engines still make gender, number and dialect errors and mishandle mixed right-to-left lines.
Can Arabic MTPE target a dialect?
Arabic MTPE usually targets Modern Standard Arabic, because most engines are trained mainly on MSA. Content meant to read in Gulf, Egyptian or Levantine Arabic needs heavier editing. For marketing or conversational text, human translation by a native speaker of that dialect is often the better choice. El Turco agrees the variant per market at scoping.
Can LSPs outsource Arabic MTPE to El Turco?
LSPs can outsource Arabic MTPE to El Turco as a regional vendor. Native post-editors work under ISO 18587 in the CAT tool or portal the LSP uses, to the agreed glossary, style guide, target variant and post-editing level. Every job passes What's Wrong? and a human sign-off, and the LSP can request LQA sampling on top.
How long does Arabic MTPE take?
Turnaround is agreed at scoping, with a 24-hour option for small urgent updates. Volume, the post-editing level and the number of review steps decide the schedule for larger jobs. The account manager confirms capacity before a job is accepted, so every Arabic MTPE deadline is agreed rather than assumed.
Ready to scope Arabic machine translation post-editing?
LSPs can send a free test of up to 300 words. Enterprises can start with a free pilot on their own Arabic content. MTPE is quoted after scoping and a sample of the engine output.
A project manager reads your brief, usually the same working day.