Warp Unveils Code Diff-Tracking Tools to Enhance AI Coding Transparency and Control
Warp is introducing a new suite of tools aimed at making AI-powered coding more transparent and collaborative, positioning its approach as a modern take on pair programming. The company unveiled Warp Code, a set of features designed to give developers greater visibility and control over AI coding agents operating in the command line. At the heart of the update is enhanced diff tracking, allowing users to see every incremental change the AI agent makes in real time. “With other command-line tools, you’re just crossing your fingers, hoping the output is something you can actually merge,” says Warp founder Zach Lloyd. “Our goal is to create a much tighter feedback loop for this agentic style of coding.” Now, as the AI writes code, users can view each individual change—what’s been added, modified, or removed—step by step. They can also comment directly on specific lines, ask questions, or adjust the agent’s behavior mid-process. This interactive layer turns the coding experience into a dialogue rather than a one-way instruction. The interface remains familiar to existing Warp users: a bottom panel for inputting commands, a central window showing the agent’s responses, and a side panel that displays a detailed, time-stamped history of code changes. Users can edit code manually, just like in tools such as Cursor, but they can also highlight specific sections to provide context or ask follow-up questions. One standout feature is the built-in compiler, which automatically detects and troubleshoots errors during compilation, reducing friction in the development workflow. “This is about ensuring you understand what the agent is producing and giving you the ability to review, refine, and guide it,” Lloyd explains. Warp is entering a crowded space, competing with AI coding tools like Cursor and Windsurf, as well as non-code platforms such as Loveable. It also faces indirect competition from foundational model providers—Anthropic with Claude Code and OpenAI with Codex—whose models Warp leverages to power its own product. Despite its relatively small size, Warp is growing quickly, with over 600,000 active users. Lloyd reports the company is adding $1 million in annual recurring revenue every 10 days, signaling strong demand for a more intuitive, transparent way to work with AI in code. The new features aim to solidify Warp’s position as a tool that doesn’t just write code—but helps developers understand and shape it along the way.
