CLOVA AI Tech Demo

CLOVA AI envisions AI for Everyone.
We aim for world's best AI technologies and
want to provide meaningful services.

CLOVA AI is an artificial intelligence platform created by Korea’s No. 1 search engine, NAVER, and Japan’s No. 1 messenger, LINE. The world’s top-level artificial intelligence technologies NAVER and LINE have been studying for around 20 years, such as search, speech recognition and synthesis, image and video analysis, natural language processing and chat functions based on numerous contents and services of NAVER and LINE are provided on a single platform. In addition, by using CLOVA platform, everyone can conveniently create artificial intelligence services. CLOVA AI is making an AI ecosystem and a more convenient world together with the users and a wide range of partners.

  • Speech Recognition

    CLOVA Speech is a world-class speech-to-text system providing robust recognition performance under noisy conditions for Korean and Japanese languages.

  • Speech Synthesis

    CLOVA Speech Synthesis technology is the best in the world. We creates the most realistic artificial voice that sounds like celebs or your friends. Your text becomes the most natural voice with CLOVA Speech Synthesis.

  • Smart Lens

    CLOVA Vision enables users to conveniently search various objects visible to the eyes in everyday life through images. CLOVA Vision technology has been applied to Line Shopping Lens, Naver Smart Lens, and Naver Shopping Lens to provide a new search experience and replace the existing text-based search services. The performance has been optimized through the largest search data in Korea.

  • Video Highlight

    CLOVA Video Highlight is an AI technology to enable capturing only the selected person. It is capable of selecting video segments where only the selected person is displayed and even when the face is hidden. The accuracy of Video Highlight’s capturing function is 99% given the entertainers’ performance videos. The performance has been proved through V LIVE feature where videos of only specific people can be edited.

  • Watermarking image Watermarking

    CLOVA Watermarking is a technology to embed specific information in an image for intellectual property protection. Although a Marker that is invisible to human eye is applied, the Marker is protected even after the video goes through various conversion processes. The accuracy of CLOVA Watermarking technology is 98%. It can be applied right away to protect intellectual property rights when distributing images and video contents, such as WEBTOON.

  • OCR image OCR

    OCR is the technology to find the location of letters in a photo and identify the type of letters. CLOVA OCR uses an independently developed letter field detection technology to find and identify letters of various languages and styles. It has been ranked No. 1 in four OCR fields’ ICDAR leader boards and can be applied to various services including translation of foreign languages in images as well as recognition of Japanese or Korean receipts and business cards.

  • NSML

    NSML is a cloud machine learning platform that ensures convenient and swift R&D in the deep learning algorithm design process. It provides automatic GPU scheduling and AutoML function for automatic hyper-parameter search. In addition, it enables users to collaborate with coworkers through a leader board and applies deep learning models as the API format to a service right away. It has already been proved through API provision in CLOVA.

  • Face image Face

    It provides the entire pipeline technologies from face detection to recognition. As it does not require learning for an additional face, a new face can be quickly registered and recognized. Through the independently developed weight reduction technology, real-time recognition is possible even in the CPU and mobile devices. This technology can also be applied at a low cost to large-sized data, such as videos. It is a core function for face recognition in videos such as V LIVE.

  • Photo2Art

    This is the technology to convert a user’s photo into a work of art. Photo2Sticker search function recommends a cute Line sticker that resembles the user’s selfie. Photo2Art creation technology is a world’s top-level technology to more realistically, magnificently and artistically convert users’ photos to desired styles on a real-time basis.

  • Chatbot image Chatbot

    ​With CLOVA Chatbot Builder, you can easily create Chatbots that understand the intent of user’s question and give the right answer. The powerful dialog engine based on natural language processing technology and machine learning algorithm supports not only Japanese but also various languages such as English and Korean. It can be applied to a variety of services such as customer inquiry, ordering, and reservation. Currently, CLOVA Chatbot Builder is used in various services such as Line, Naver and LG U+.

Contact Us

The vision of CLOVA AI is to achieve "AI for Everyone”.
For any inquiries or questions regarding our technology and partnership, please contact us at for more details.