Speed and Efficiency:
MTPE accelerates the translation process by quickly processing large volumes of text through a trained engine, followed by human revisions and refinement.
Reduce translation costs by up to 50% and speed up turnaround times.
Machine Translation Post-Editing (MTPE) combines machine translation with expert human editing for fast and accurate multi-language content translation. The merger of technological efficiency with linguistic expertise offers businesses quick turnarounds, a broad language range, cost savings, and various levels of edit to make sure the quality meets your program goals.
I've worked with Alconost for several years now, and I've always been impressed with their professionalism and attention to detail. They consistently deliver high-quality translations that are accurate and culturally appropriate. I highly recommend their services to anyone looking for a reliable localization partner."
We have localized over 1,500 projects and produced 1,000+ videos. Let's make something special for you too!
A GenAI or NMT (Neural Machine Translation) engine that fits your project
Dedicated project manager
A GenAI or NMT (Neural Machine Translation) engine that fits your project
Dedicated project manager
Essential editing made by professional linguists to spot inaccuracies
Engine evaluation
A GenAI or NMT (Neural Machine Translation) engine that fits your project
Dedicated project manager
Essential editing made by professional linguists to spot inaccuracies
Meticulous editing made by professional linguists to ensure high-quality, fluent, and contextually correct content.
Access to the Localization Management Platform
Engine evaluation
Everything in Full Post-Editing, plus:
Personalized service for custom needs, such as engine training, terminology management, and meeting other specific requirements.
During the preparation stage, the project manager (PM) reviews the content and collects additional context if required (e.g., type of content, audience, tone of voice, quality expectations, etc.). In the preparation stage, the PM:
Selects a sample from the text (typically ~2000 chars, up to 2 languages) that will be used in engine evaluation.
Pre-selects 7-9 Neural Machine Translation (NMT) engines or GenAI for evaluation.
(Optional) Conducts prompt engineering if GenAI is used.
Sets up a test project on the localization platform.
The PM performs the following actions to determine which engine(s) perform the best for the given project. There are 2 types of evaluation Express and Deep:
Pre-translates the text into all project languages using the pre-selected NMT engine or GenAI. Express Deep
Sends the output translation samples to native-speaking human linguists for evaluation. At this stage, the translation outputs generated by each NMT engine or GenAI are reviewed and holistically evaluated. Express Deep
Determines the translation sample with the highest evaluation score, which will then be post-edited by human translators. Deep
Requests engine scoring. Each translation sample is assigned a score derived based on industry-accepted metrics used to compare sample translations against the post-edited translation (COMET, CHRF, BLEU, hLEPOR, WER). Engines with the highest scores will be used in the project going forward. Based on our experience, NMT/GenAI perform differently for different language pairs and types of content. Deep
Based on the holistic evaluation and metrics above (in case of Deep evaluation), the PM prepares the project estimates (budget, turnaround time).
The project manager builds the localization workflow, pre-translates content using the NMT/GenAI selected above, allocates tasks to linguists on the localization platform, and monitors deadlines and translation quality.
If translators have questions, they ask them on the localization platform. Discussions are visible to the whole project team.
The completed translations are automatically exported in the same format as the original files using an API or manually.
The PM also conducts recurring QA checks to validate that the output quality is continuously high and that the NMT/GenAI perform as expected. Depending on the project requirements, the NMT/GenAI can be regularly retrained based on glossaries and translation memory accumulated during the project.
Receive immediate raw translation output produced solely by advanced machine translation engines, without any human intervention. Benefit from a Managed Service, supported by a dedicated project manager who ensures that your localization process runs smoothly from start to finish.
It’s best for large-scale projects demanding fast and efficient translation, such as:
Build on the advantages of RAW MT with essentials-only editing by professional linguists to render the translated content comprehensible, addressing any inaccuracies that could misconstrue meaning or trigger confusion.
For non business-critical content, including:
A comprehensive review and correction of machine-generated translations by professional native linguists. The objective is to produce a final product that is the same quality as a human translation, with proper attention to style, tone, and context or culture-specific nuances.
Track process and progress to ensure high-quality, fluent, and contextually correct content with access to our Localization Management Platform.
For high-volume urgent tasks and mission-critical content, such as:
Request a unique solution for your specific project needs, integrating advanced MT with customized workflows and human expertise.
We will tailor your translation service to match the nuances of your content. Pricing depends on your particular needs.
This is a comprehensive package for solutions like:
When deciding between Machine Translation Post-Editing (MTPE) and traditional human translation, several factors come into play. MTPE offers notable advantages over traditional methods:
MTPE accelerates the translation process by quickly processing large volumes of text through a trained engine, followed by human revisions and refinement.
With less intensive human input when compared to traditional translation, MTPE often leads to lower costs while leveraging advanced machine translation technology.
MTPE scales well for high-volume multilingual projects, quickly providing base translations across various languages for subsequent post-editing.
Machine translation maintains consistency in terminology and style, which is particularly beneficial for large volumes of text, such as in traditional localization. Human editors then refine it to match specific guidelines.
Choose MTPE for projects prioritizing speed, scalability, and cost-effectiveness. Quality is guaranteed by the post-editing process, in which human linguists review content to make sure the quality meets expectations. Traditional translation remains ideal for projects that focus on detailed, high-quality content, demanding deep cultural understanding, like creative and marketing materials and sensitive or confidential documents that require human accuracy and security.
MTPE begins with an automated translation which is generated by a machine translation (MT) engine. The raw translation is then sent to a specialist linguist called a post-editor, who reviews and checks the translation for accuracy and consistency. This may involve making corrections or amending the translation in order to make it as accurate and clear as possible.
There are several Machine Translation Post-Editing (MTPE) options and it is important to choose the best fit for your business. To simplify your decision-making process, we've outlined 4 key considerations to help you select the most suitable package for your project:
Raw MT output may or may not meet your requirements for quality. The level of translation quality you need depends on the visibility, level of nuance, and importance of your content. For example, user generated content is neither sensitive nor of critical importance so is a viable candidate for MT + light post-editing. However, website content is of high importance, highly visible, and often has an emotional nuance. The more sensitive to quality the content is, the more advanced MTPE package you'll need to ensure a polished, error-free translation.
More advanced packages provide more thorough editing that may take more time in comparison to simpler ones.
Some languages are more viable for MT processes than others because machine translation engines are not yet well-trained enough to accurately translate content into them. This results in the need for more careful post-editing for those language pairs.
If your content deals with sensitive information, requires a nuanced understanding, or contains humor, then you should consider more advanced packages. It’s this advanced level of post-editing that ensures every translated word aligns perfectly with your original intended message.
Every project is unique, and so are its translation needs. Depending on the project, you may combine different MTPE packages. For instance, you might blend Full MTPE for critical content with RAW MT for support materials. We will work with you to tailor the solution to suit the specific needs of your project.
Engine Evaluation is the process of assessing the performance and accuracy of various machine translation engines. It involves testing and analyzing how well the engine translates different content types and evaluating its strengths and weaknesses. This evaluation helps determine the suitability of the engine for specific translation tasks, languages, or content types and allows for optimization or customization to improve translation quality.
MTPE, or Machine Translation Post-Editing, is often more cost-effective than traditional translation services because it combines the speed and efficiency of machine translation with the expertise of human editors. Additionally, MT is quicker, and depending on the level of post-editing required timelines can be substantially condensed. However, it’s important to note that MT is not free: customizing and fine-tuning the engine bears a cost.
Generative AI (Large Language Models) excels at creating new content by learning from diverse data, ideal for tasks that need innovation and cultural adaptability. NMT, however, is focused on accurately translating text from one language to another and is better suited for industries where precision is critical, like healthcare or law. Regardless of the technology, integrating human oversight can boost translation quality. Choose Generative AI for creative flexibility or NMT for precise translation, based on your project's requirements.
All translations at Alconost are performed by professional translators who are native speakers.
We have approximately 700 professional native-speaking translators on our team, who specialize in a broad range of fields. We always try to assign one or more dedicated translators to your project and work specifically with these translators on a continuous basis. This is useful if you have a large project that is updated periodically, you need to use consistent terminology, or you simply like a particular translator's style. If the translator changes for some reason, they will provide the glossary and translation memory to the new translator to ensure consistency of terminology and style in future updates.
Our first choice is the cloud-based translation platform Crowdin. It allows us to manage localization projects in real-time. You can:
We also work on other localization platforms, such as GitLocalize, Phrase, Lokalise, Memsource, memoQ, Smartcat, WebTranslateIt, Transifex, OneSky, POEditor, Trados, and many more.
We can work with any file type including:
Android .xml, iOS .strings, .stringsdict, Windows .resx, .po, .pot, .php, .json, .ini, .csv, .js, .plist, .xlf, .xliff, .stf, .dita, .ditamap, .toml, .gotext.json, .arb, .vdf, .wxl, .nsh, .properties, .dtd, .ts, .rc, .resw, .resjson, .yml, .htm, .html, .xhtml, .haml, .txt, .md, .xht, .idml, .docx, .svg, .pdf, .xaml, .srt, .vtt, .sbv, .mif, .idml, .rtf, .ppt, .odt, .ods, .odg, .odp, .wiki, .flsnp, .flpgl, .fltoc, .md, .xht.
We also work with other file formats and proprietary systems. Let's discuss your project!
We translate into more than 100 languages, including the most popular (English, Chinese, Japanese, German, French, Spanish), and rarer ones, and even dialects of one language such as Spanish for Spain versus Spanish for Latin America.
We can find a translator in almost any language that you need, depending on your product and goals. Sometimes localization into niche languages can offer better prospects than localization into mainstream languages.
To help us calculate the cost of localization, please request a quote with the target languages that you would like to localize your website into.
We have localized over 1,500 projects and produced 1,000+ videos. Let's make something special for you too!
Have a project in mind?
We’d like to learn more about it. In return, we’ll get back to you with a solution and a quote.
Looking for a new job? Apply here!