Sichuan Cooking Robot Recreates Wok Hei with Sensors, Vision and 10,000 Recipes
Robotic cooking is no longer a novelty, but a robot that can produce wok hei is still striking. Recently, a Sichuan-made cooking robot drew attention for simulating a chef’s tossing motions and preserving the distinctive wok hei of Chinese dishes.
Wok hei refers to the volatile aromatic compounds created when ingredients undergo Maillard reactions and caramelization in a very hot iron wok. It is the smoky, seared aroma that gives many Chinese stir-fries their character. Whether a dish has wok hei often depends on a chef’s on-the-spot judgment and accumulated experience. For a cyber chef, the real challenge is not copying a recipe but replicating that judgment.
According to Huang Tianyong, founder of Sichuan Xiong Miao Master Technology Co., the old assumption that robotic cooking is only an execution problem no longer holds. He says it is also a problem of understanding. A robot must not merely memorize steps and execute them; it must decide when to add ingredients, when to reduce the sauce, and when to turn up the heat. Those decisions depend on the state of the ingredients, the heat in the wok, the color of the sauce, and the degree of stirring.
That shift matters because cooking is full of variables. A fixed program cannot respond to them; an understanding system can.
Observation comes first. The same cut of meat can differ in size, freshness, and initial temperature; the same vegetable can vary in water content. An experienced chef observes and handles these variables. The cooking robot has two kinds of eyes: sensors inside the wok judge whether ingredients are cooked and whether the sauce has reduced properly, while a vision system checks whether the dish’s color is right, helping decide whether to raise or lower the heat. These two inputs replace a chef’s sight and touch, letting the machine make real-time calls rather than follow a fixed script.
Huang says heat control is the key to giving food wok hei. Wok hei often depends on fierce, rapid cooking, and some Chinese stir-fries are won or lost in an instant. If the fire is too strong, the dish burns; if it is too weak, the dish releases water and becomes soft and unpalatable. Take stir-fried pork kidney: it often must leave the wok within seconds. One second less can leave it undercooked, one second more can make it tough. With a smart brain, the cooking robot can dynamically adjust heating power, cooking duration, stirring speed, and the amount of seasoning. The window for perfection is extremely narrow, which is why experience matters so much; the robot’s brain must continuously balance these parameters.
With eyes and a brain, the cooking robot has the potential to be a chef, but becoming a true chef requires understanding cooking. Huang says good taste is not only about precision; it should also carry emotional resonance and memory. The company spent several years collecting more than 10,000 recipes from China’s eight major cuisines, decomposing and encoding chefs’ cooking motions and processes for the robot to learn. The database covers Sichuan, Cantonese, Shandong, Huaiyang, and other major cuisines, giving the robot a broad culinary foundation. The goal is to let the machine absorb not just ingredient lists and times, but the logic behind a chef’s actions.
It has also developed systems such as a large model, allowing users to customize a personal recipe with a single sentence and have the robot prepare it. A user trying to lose weight could get a low-sugar, low-salt version of braised pork, while someone who has long lived away from home could taste the flavors of their hometown. Such customization points to a future in which cooking robots are not just kitchen appliances but platforms for personalized nutrition and nostalgia.