Woof woof! Can Chinese AI pet collar truly help you 'hear' your pet?

Your pet cats and dogs may soon get what the 'hooman' is saying. And, the hooman may also be wiser of what their four-legged wards are barking or meowing about.

A China-based startup has developed an AI-powered smart collar that claims to translate pet sounds into human language with up to 95 per cent accuracy. Called PettiChat, the wearable device has been developed by Hangzhou-based startup Meng Xiaoyi and is powered by Alibaba Cloud’s Qwen AI model.

The device has already generated significant online buzz, reportedly crossing 10,000 preorders ahead of its official May 30 release. Preorders opened on May 1.

What's in that collar?

Priced at 799 yuan (roughly Rs 11,300 or USD 118), the smart collar can be attached to a pet’s existing collar. It uses built-in microphones, motion sensors and accelerometers to capture animal sounds, movements and behavioural patterns.

Also read: Siberian husky tragedy in Telangana exposes India's cruel exotic pet culture

According to the company, the system analyses these inputs using a hybrid device-cloud AI model trained on millions of pet sound samples and matched video clips.

Meng Xiaoyi claims the collar can identify more than 20 emotional expressions and convert them into human-readable phrases within 1.2 seconds. Demonstration videos released by the company show a cat meowing while the app translates it as “I wanna play,” while a barking dog is interpreted as saying, “I’m hungry.”

One of the device’s most talked-about features is its claimed two-way communication system. The company says owners can speak commands such as “Easy, stay calm,” and the collar will convert them into pet-friendly audio signals or cues.

The collar also includes GPS tracking, smartphone connectivity and geo-fencing alerts to notify owners if pets wander beyond a designated area.

Where's the proof?

However, the product has also sparked scepticism online. Critics argue that the startup has not published any independent, peer-reviewed studies or transparent datasets to support their 94.6–95 per cent accuracy rate claim.

Also read: Explainer: How best pet owners can handle their 'ferocious' dogs

Several users on social media dismissed the device as an “IQ tax,” claiming the AI likely relies on predictive mood classification rather than true cross-species translation.

Researchers and veterinary experts have similarly questioned whether the system can reliably interpret complex emotional states outside controlled testing environments.

Critics also noted that Meng Xiaoyi was founded only in January this year, although the company has reportedly already secured USD 1 million in seed funding.

The first, really?

The idea of translating animal sounds into human-understandable language is not entirely new. Japan’s Takara launched BowLingual in 2002, widely considered the world’s first commercial pet translator. The gadget classified dog barks into basic emotional categories using a wireless collar and handheld receiver. Later, smartphone apps such as MeowTalk attempted similar mood-based translations for cats.

Also read: International Dog Day: Want a furry friend? Here are some breeds suited for India

More recently, AI-powered pet communication tools have expanded with devices like the Traini AI Collar showcased at CES 2026, focusing on converting human speech into dog-friendly acoustic signals.

What sets the latest device apart, according to the company, is the use of generative AI. Unlike earlier devices that depended on fixed databases of pre-programmed phrases, PettiChat uses large language models to generate more natural and variable responses in real time while combining sound analysis with motion-tracking data.

Woof to that.