For Platform 2.0 Only

Papago NMT

입력한 텍스트를 인공신경망 기반 번역 알고리즘을 통해 여러 나라의 언어로 자동 번역해줍니다

The Artificial Neural Machine Translation API Providing the Most Outstanding Korean Translation Quality

This translation service uses Neural Machine Translation (NMT) based on the artificial neural network-based machine translation technology developed by Naver.
Because this technology translates the phrase as a whole without breaking the input sentence down into segments, a context-aware translation can be obtained which is more natural than other machine translations.

Rich Korean Language Processing Experience
More accurate and natural translations can be achieved by applying NAVER's artificial NMT engine that processes multiple languages based on Korean. This technology has currently been adopted by a variety of services, such as the Papago Translation app, NAVER Dictionary, NAVER SmartBoard, NAVER i, Clova, and the LINE Translation Bot.
Performance Improvement through Learning
To strengthen the translation engine technology which effectively identifies the language rules, continuous learning is performed with a massive amount of data owned by Naver services. Improve the quality of your services with Papago NMT, the service that is building on the abundant data of Naver to provide faster, more natural translations.
Safe Personal Information Management
To safely protect the information entered for text translation, Papago NMT service complies with Korea's privacy laws, such as the Personal Information Protection Act and the Act on Promotion of Information and Communications Network Utilization and Information Protection, etc.
Easy Use
On the web-based console, you can conveniently handle various types of management tasks, such as service registration (register, change, and delete), setup of usage limits (daily/monthly), and checking usage statistics.
Various Additional Features
In addition to the translation service, we provide a language detection service that automatically detects the type of text and language to be translated. Additional features are also provided, such as the language detection service that automatically recognizes the source language without the need for the user to specify it.

Detailed Features

A high quality machine translation service that learns from NAVER's abundant data is provided.

Overview of NMT Technology

Neural Machine Translation (NMT) is a translation technology that uses an enormous virtual space called an artificial neural network to understand and translate sentences.
Input sentences for translation are converted to vectors (coordinates) that represent specific points of the artificial neural network in the virtual space and then saved. Sentences with similar meanings become clustered together. NMT references the information surrounding the input sentence to understand its context for a more natural translation.
Papago NMT self-studies the large-scale data accumulated through NAVER services every day. Through this translation model, translation results that are more accurate than the existing SMT method can be provided.

Provided Services

Translation API (Papago NMT API)

When the source for translation, code for the language to translate into, and code for the translated language are delivered to the Papago engine, this API service returns the translated result.
Translation of 16 languages pair are possible.
Korean, English, Chinese(Simplified/Traditional), Japanese, French, Spanish, Vietnamese, Thai, Indonesian, German, Russian, Italian, etc.

Papago Language Detection API

This API automatically detects the language of the source text for translation. This is a learning-based language detection model developed based on a massive language corpus. A fast and accurate language detection feature is supported for the 12 main languages.
Translation of 12 languages are possible.
Korean, English, Chinese(Simplified/Traditional), Japanese, French, Spanish, Vietnamese, Thai, Indonesian, German, Russian, Italian, etc.

Using the API

As an open API that does not require a login, service usage is possible by sending the Client ID and the Client Secret value issued when you registered the application in the HHTP header at the time of calling POST.

Use Cases

This technology is adopted by a variety of services, such as the Papago Translation app, NAVER Dictionary, NAVER SmartBoard, NAVER i, Clova, and LINE Translation Bot.

Notes for Use

  • Papago NMT currently provides a short-sentence based service that returns a translation of the user’s input text.
  • Translations that include HTML tags are referred as 'web translation' and this feature is currently not supported.
  • The maximum number of characters you can enter at a time is 5,000. (including spaces)

Service Usage Limit

  • You can directly adjust the usage limit of the Papago NMT API and the Papago Language Detection API from the service console.
  • You can adjust the usage limit of Papago SMT yourself using various options from 3 million characters a month (100,000 characters a day) to 300 million characters a month (100 million characters a day) per application.
  • Calling is allowed for the last call that exceeds the usage limit. (The usage fee is calculated based on the number of actual calls.)
    E.g., you have set the monthly usage limit to 1,000 characters. Once you use 990 characters, you can call an extra 20 characters. Then, you can use a total of 1010 characters and will be billed for 1010 characters. 
  • If you need to increase the service limit, please contactCustomer Support.

Pricing Information

The Papago NMT usage fee is charged in 1,000,000 character units based on the requested number of characters.

ServiceCharge (Monthly)Billing UnitsUsage Fee
Papago NMT APINumber of characters entered for translation1,000,000 characters20,000 KRW/1,000,000 characters
Papago Language Detection APINumber of characters entered for language detection1,000,000 characters20,000 KRW/1,000,000 characters
  • A service fee is charged every 1,000,000 characters. Therefore, the number of characters input for translation is rounded up to the nearest 1,000,000 characters for calculation.
    (E.g., If you use 500,000 characters, then you will be charged 20,000 KRW and if you use 1,600,000 characters, then you will be charged 40,000 KRW.)
  • The character count includes spaces.
  • The usage (number of characters) is calculated on a monthly basis and the remaining characters are not carried forward to the next month.
  • The usage of Papago NMT API and Papago Language Detection API is counted separately.
    (E.g., If 500,000 characters in the Papago NMT API are processed in a month, plus an additional 300,000 characters are processed in the Papago Language Detection API in a month → 20,000 KRW + 20,000 KRW → 40,000 KRW will be charged.)

Was this page helpful?

Please share your opinion and any suggestions for us.
Please enter content.
Send Opinion