DoorDash launches command-line tool for AI agents to order food

DoorDash has rolled out a beta version of its command-line interface tool for developers, opening up the possibility for AI agents to search for restaurants, discover promotions, and complete orders without ever visiting the app. Called "dd-cli," the limited beta is available to macOS developers in the U.S. and Canada through a waitlist, according to an announcement from co-founder and Chief Technology Officer Andy Fang.

On its surface, the move seems more comedic than practical—the idea of automating lunch orders through terminal commands feels both absurd and somewhat over-engineered. Yet the initiative represents something more meaningful: a serious bet on what industry observers call agentic commerce, where software agents autonomously handle transactions on behalf of users.

Opening DoorDash to developer tools#

By exposing its ordering infrastructure through a command-line interface, DoorDash is allowing developers to embed food delivery capabilities into custom applications and services. Rather than directing users to DoorDash's native app, developers can now craft their own ordering experiences, combine food delivery with other tools, or integrate it into broader agent workflows. This approach mirrors patterns already emerging across the e-commerce and AI landscape.

A broader strategy for agent-driven ordering#

DoorDash has been testing multiple channels for agent-based commerce. The company recently introduced iMessage ordering and deployed "Ask DoorDash," its own AI chatbot that accepts natural language prompts and even photo inputs. The platform has also enabled ordering through third-party AI assistants, including OpenAI's ChatGPT and Claude, demonstrating a deliberate expansion beyond traditional digital interfaces.

The humor and the substance#

The announcement cannily references the famous XKCD comic about a programmer who automates a simple task by using elevated system privileges—"sudo make me a sandwich." The accompanying video highlights the humorous complexity of the feat: the system reads Slack messages, retrieves stored memories, parses JSON, examines menu data, runs Python routines, handles errors, and calculates totals, all to order three salads. This tongue-in-cheek presentation masks a genuine strategic move to position DoorDash at the intersection of developer tools and AI-driven commerce.