Xinapse’s AI virtual voices reproduce the voice of deceased former Korean president Park Chung-hee delivering messages to future generations through JTBC's 20th presidential election counting broadcast. — Photo courtesy of the company
Korean artificial intelligence (AI) startup Xinapse announced its launch in Viet Nam on November 29 by establishing a joint venture and outlined plans to expand further into the market.
Xinapse was the first to commercialise virtual voice and AI synthetic voice services to the market. The company is recognised as having high-quality speech synthesis technology and AI communication technology in South Korea. It has launched the very first virtual voice navigation service called 'T-map Celeb', which utilises text-to-speech (TTS) technology where an AI voice generator reads text out loud in a humanlike voice. This enables drivers to hear popular celebrities in voice-guided car navigation systems.
Through generative AI, Xinapse develops its own technology, such as NLU (Natural Language Understanding) and the ‘DeepSpeech’ speech engine as well as a video engine named ‘Live portrait’. With these solutions, B2B customers can create original text, images, and videos based on simple human input. The company’s AI-based solutions are currently integrated in various fields, such as TV broadcasting, audiobooks, advertisement, metaverse, contact centre, public sector and more.
"We're delighted to establish Xinapse Vietnam to tap into this growing market. Our global expansion is a key step in proving our competitiveness by reaching a broader customer base. We continuously strive to optimise our solutions to best meet the growing demand in various fields in Viet Nam,” said Dong-Won Joo, CEO of Xinapse.
In its expansion phase, backed by a joint venture established with a local partner, the company seeks to combine various AI engine applications to create use cases in Viet Nam. Also, the company aims to develop and scale local Vietnamese speech data while providing job opportunities for Vietnamese people.
In addition, Xinapse offers an image data labelling solution that can analyze unstructured big data into quality image data. The solution is designed to boost efficiency in analyzing a large amount of image data as well as save time. Customers can easily label and automate large amounts of data and customize the algorithm tailored to their needs.
Founded in 2017, Xinapse specialises in generative AI-based solutions, mainly conversation engines to generate chatbots, speech engines to generate virtual voice, and video engines to create synthetic content to enrich communication experiences. — VNS