In mid-November, WeChat quietly introduced voice-to-text support for the Teochew dialect, marking the second Chinese dialect added after Cantonese. Users who update to the latest version can now press and hold the microphone button in chat windows, swipe up to convert speech directly into text, or long-press voice messages to transcribe them—enabling bidirectional translation between Teochew and Mandarin.
Many users have shared their experiences on social media, testing phrases like "Have you eaten?" or "What are you doing?" in Teochew and praising the feature. Some joked that the "encrypted communication" of niche dialects has now been "cracked." However, WeChat’s recognition accuracy for longer sentences, slang, or heavily accented Teochew remains imperfect.
Automatic Speech Recognition (ASR) is a system that converts human speech into text or commands through feature extraction, pattern matching, and model training. Native Teochew ASR support is seen as part of WeChat’s strategy to enhance engagement among specific, high-demand user groups as its growth matures.
Launched in January 2011, WeChat has grown into a national super-app over 14 years. Tencent’s Q3 earnings report shows combined monthly active users (MAU) for WeChat and Weixin reached 1.414 billion. While WeChat first disclosed MAU figures in 2013 (272 million in Q3, up 124% year-over-year), growth has slowed significantly since surpassing 1 billion MAU in 2018. By 2025, year-over-year MAU growth dipped to around 2%, with quarterly growth as low as 0.2%—down from 0.8% in the same period last year.
Teochew, a dialect with over 10 million speakers globally, is deeply rooted in the Chaoshan region (Shantou, Chaozhou, and Jieyang), where the combined resident population hit 13.835 million in 2024. The dialect also thrives among diaspora communities nationwide and overseas.
WeChat’s ASR development from Mandarin to Teochew spanned six years. The feature debuted in June 2019 with Mandarin support, added Cantonese in September 2020, and only extended to Teochew in November 2025.
Classified as a Southern Min dialect, Teochew is considered a "low-resource language" in AI and speech recognition due to data scarcity, complex tones (eight distinct tones), and accent variations, making ASR development challenging. However, it isn’t the second-largest Chinese dialect after Cantonese. Ethnologue data shows Mandarin remains the dominant native language (941 million speakers in 2024), followed by Cantonese and Wu (over 80 million each), then Southern Min (74.02 million). Hakka, Jin, Xiang, and Gan dialects have 48.2 million, 47 million, 37.3 million, and 22.1 million speakers, respectively.
Comments