Agility’s Digit Robot Follows Natural Language Command to Deliver Pasta Ingredients
Agility Robotics has taken a major leap toward practical, household humanoids with a new demo showcasing its flagship robot, Digit, executing a multi-step task entirely from a natural language prompt. The command was simple: “Bring me the ingredients to make pasta.” The execution, however, was anything but.
In the demo, Digit autonomously interpreted the request, navigated its environment, identified relevant objects, and delivered a collection of pasta-making ingredients—such as pasta noodles, tomato sauce, garlic, and olive oil—to the user. This demonstration highlights a major breakthrough in human-robot interaction, integrating natural language processing, visual recognition, and robotic manipulation in a cohesive workflow.
Multi-Step Reasoning in Action
Unlike rigidly pre-programmed robots, Digit understood the intent behind the command and made autonomous decisions about what items were required and where to find them. Using embedded AI and real-time perception systems, it:
Parsed the user's spoken request using a built-in large language model.
Cross-referenced the task with a knowledge base about common pasta recipes.
Located the relevant items using object recognition.
Planned and executed a retrieval path, navigating obstacles in a real-world kitchen setting.
Delivered the items to the user without requiring any further prompts.
A Glimpse into the Future of Home Robotics
This demonstration marks a turning point for household and service robots, especially in environments where adaptability and communication are key. While robots have long been capable of performing isolated tasks, the ability to understand context, chain actions, and interact with humans naturally signals a new era.
Agility Robotics envisions Digit as a flexible platform for both logistics and domestic environments. Initially focused on warehouse and industrial use cases, Agility is now testing how humanoid robots can assist in homes, restaurants, and eldercare settings—where intuitive interfaces and adaptive behavior are essential.
Why It Matters
Natural language task execution isn’t just a novelty—it’s the cornerstone of making robots truly helpful in everyday life. From assisting the elderly to streamlining home chores, the ability to ask a robot for help in plain English is what will make robotic assistants mainstream.
Digit’s successful pasta run might seem small, but it's a big step toward a world where housebots are as commonplace as dishwashers and voice assistants.
