Weverse Brain Commercializes Its Proprietary TTS Engine for the First Time

by Choi Hokyung

Published 29 Apr.2026 09:44(KST)

Initially Applied to "MaxAI"

Weverse Brain announced on April 29 that it has commercialized its self-developed TTS engine in "MaxAI."

The newly developed TTS engine has been initially applied to select features of the company’s conversational AI tutor service, "MaxAI." With the completion of both its previously developed STT engine and now the TTS engine, Weverse Brain has established full-stack voice processing technology for conversational AI.

Weverse Brain logo. Weverse Brain

The core features of the proprietary TTS engine are threefold: ▲ speech quality for non-English languages, ▲ processing of mixed bilingual utterances (code-switching), and ▲ utterance instruction control technology.

Speech quality for non-English languages is an area in which global commercial engines are relatively weak. Weverse Brain’s technology minimizes pronunciation and intonation errors, especially in character-based languages. The company’s accumulated phoneme analysis and pronunciation understanding for each language-built through its multilingual education services-serve as the foundation for correcting such errors.

Mixed bilingual utterances refer to the technology that seamlessly processes language switching within a single sentence where different languages are mixed. Unlike English-speaking regions, Asian language environments often naturally blend foreign languages into everyday conversation. For example, in the sentence "The deadline is by 8 o'clock," there may be cases where "dead" is pronounced in English and "line" is pronounced awkwardly in Korean. Weverse Brain’s proprietary engine is designed to handle such points of language transition naturally.

Utterance instruction control is a technology that specifies detailed elements of speech such as intonation, stress, speed, and tone. It enables the design of speaking styles tailored to specific purposes, such as emphasizing key sentences or delivering information slowly. This functionality is expected to be widely applicable not only in educational settings but also in AI counseling, sales response, content narration, and various other conversational AI services where effective delivery is critical.

Sewon Cho, CEO of Weverse Brain, stated, "The internalization of TTS technology is a foundational step toward creating an environment where AI speaks as naturally as humans, beyond just language learning." He added, "We will continue to advance core technologies for conversational AI and expand into a variety of business fields."