ElevTalk Closes Pre-Seed Round for Voice Interface Designed for Non-Native English Speakers

ElevTalk, a voice interface platform designed for non-native English speakers in the workplace, announced on Wednesday that it has closed a pre-seed funding round.

The amount was not disclosed.

Bass Ventures and IO Studio’s jointly managed Patriot Fund, Xiaoxiao Fund, and an angel investor who was a founding member of Baidu participated in the round.

Headquartered in Silicon Valley, ElevTalk provides a service that converts speech into text with a single tap in any app. Voice input is transcribed and refined in real time into ready-to-send text, and users can speak in their native language and have their intent translated into English through a one-touch translation feature.

Unlike existing general-purpose dictation tools for native English speakers such as Wispr Flow, ElevTalk is built for Asian professionals. The platform specializes in native transcription of English, Korean, Chinese, Japanese, and Vietnamese, natural code-switching, and enhanced handling of accents, names, and technical terminology. In its own benchmarks, ElevTalk recorded a character error rate (CER) of 1.55% for Korean and 2.25% for Chinese, outperforming major global dictation tools including Wispr Flow and Typeless. For Korean, the error rate is approximately 24% lower than Wispr Flow; for Chinese, approximately 43% lower. The company says ElevTalk is currently the only product to achieve this level of performance in both Korean and Chinese simultaneously.

Since launch, ElevTalk has acquired over 1,000 early users without marketing spend. The user base is anchored by Korean and Chinese founders and investors in Silicon Valley—early adopters with the most demanding multilingual communication needs. Core users engage actively five or more days per week, with the most active user inputting 4,000 words per day by voice. Week-over-week retention of weekly active users continues to rise from 17% to around 50%, and a significant number have switched from major global dictation tools to ElevTalk for more accurate Asian-language recognition, translation, and code-switching experience.

ElevTalk was co-founded by CEO George Dai and CTO Bob Fu. Dai, a Schwarzman Scholar at Tsinghua University who graduated from Peking University’s mathematics department, worked at McKinsey and then led AI voice strategy at China’s largest AI voice company, and previously scaled his earlier education venture Mini Gap Academy to seven-figure annual revenue (RMB). Co-founder Fu is also a mathematics graduate from Peking University who spent seven years as an early member of a leading AI startup in China, where he built products used by over one million people while gaining experience in low-latency and multimodal AI model training.

With this investment, the ElevTalk team plans to accelerate market expansion by reaching Korean, Japanese, Chinese, and Vietnamese professional communities in the U.S., and to continuously strengthen service capabilities for Asian languages and accents.

“Voice is becoming the first interface of the AI era, but general-purpose dictation products are built for native English speakers, leaving multilingual Asian professionals behind,” said George Dai, CEO of ElevTalk. “There are over 80 million non-native English speakers working in English-speaking environments worldwide, and a significant portion of them are right here in Silicon Valley. ElevTalk ensures that every Asian professional gets the most natural, professional, and ready-to-send output, no matter what language they speak, what accent they have, or how they code-switch.”

Bob Fu, CTO of ElevTalk, added, “Speech recognition in Korean, Chinese, Japanese, Vietnamese and other languages that general models fail to handle properly for Asian users, cross-lingual intent understanding, and low latency—those three things are our entire technical roadmap. Performance optimization through accent and correction feedback from actual work workflows creates a virtuous cycle that generalist players cannot easily replicate.”

Yang Hyungjoon, associate at Bass Ventures, and Kim Taeyong, CEO of IO Studio, who co-led the investment through Patriot Fund, said, “Voice keyboards are the fastest-growing interface in the world right now, but even though more than half of Silicon Valley workers are non-native speakers, general-purpose products are not built for them. We expect the ElevTalk team, which develops and improves products faster than anyone else, to grow into the business OS of the voice era, and Korea—one of the markets with the highest willingness to pay for improved English communication—will be an important anchor in that journey.”

Voice AI platforms designed for multilingual professionals are gaining investor attention as remote collaboration and global hiring expand the need for seamless cross-language communication. Companies building AI-powered voice interfaces that serve non-English-native users represent a growing segment within workplace productivity tools.