Hangzhou-based Meng Xiaoyi has developed a collar-shaped device that claims to translate cat meows and dog barks into human language with 95% accuracy.

According to vc.ru, the startup, founded in January 2026, presented the gadget at the IT Expo technology exhibition. According to the developers, the collar detects animal sounds and transmits the data to the cloud, where it is processed by Alibaba Cloud's Qwen multimodal neural network. The model analyzes vocalizations and compares them with a database of feline and canine emotion patterns. The device costs 799 yuan (approximately 118 USD). The startup has already raised 1 million USD in seed funding and collected over 10 thousand pre-orders.

However, the product has sparked a wave of skepticism on Chinese social media. Users have dubbed it a "human intelligence test," citing the lack of independent research and verified scientific data to support its claimed accuracy. Critics also noted that training a neural network to recognize animal "language" in just a few months is virtually impossible. The product's marketing campaign so far has been limited to a few short videos.

The idea of creating a translator for communicating with animals is not new. In the 2000s, Japanese engineer Kazuhiko Nakamura developed a device called Bowlingual to analyze dog barks. In 2017, Amazon patented a similar system for its Alexa voice assistant, but a commercial product was never developed. In 2024, Israeli scientists used AI to decipher bat communication. However, according to zoologists, the animals communicate not with words, but with sounds, postures, and pheromones—making the task of complete translation extremely difficult.