Skip to main content

Tibetan AI Input Method Launches, Speaks Three Dialects

A New Era for Tibetan Typing

On September 20, during a celebration of Qinghai Normal University's 70th anniversary, a groundbreaking tool was unveiled: Zhida AI Input Method, China's first Tibetan multilingual full-modal input system. Developed by the National Key Laboratory of Tibetan Intelligent Technology, this input method is set to transform how people interact with Tibetan on digital devices.

Beyond Just Typing

Zhida isn't your average keyboard. It supports text, voice, and image inputs across all devices—smartphones, computers, you name it. Whether you're typing a document, speaking a message, or snapping a photo of text, Zhida has you covered. And with a unified account, your personal dictionary and typing habits sync seamlessly across devices.

Voice and Vision

Professor Duola from Qinghai Normal University explains that the input method strictly follows national Tibetan keyboard standards, but also accepts Latin transcription, making it easier for beginners. Voice input? It understands the three main dialects: U-Tsang, Amdo, and Kham. You can even mix Tibetan, Chinese, and English in your speech. The OCR feature lets you extract text from photos or screenshots—perfect for digitizing documents or teaching materials.

Smart and Personalized

Zhida's dictionary covers terminology from 23 fields, including education, law, and literature. It learns your favorite words and expressions over time. Alongside the input method, the team released a Tibetan font set with four styles: bold, white, artistic, and brush script.

Powering the Future

The launch also highlighted the Zhida large model, which has nine versions with parameters from 800 million to 100 billion. It supports translation across 56 languages and can understand images and videos, with applications in education, research, and government affairs.

Key Points:

  • China's first Tibetan full-modal AI input method, Zhida, supports text, voice, and image inputs.
  • Covers three major Tibetan dialects for voice input and mixed Chinese-Tibetan-English OCR.
  • Includes a smart dictionary and a new Tibetan font set.
  • Backed by the Zhida large model, which handles 56 languages and multimedia understanding.