Warp Launches New Difference Tracking Tool to Boost AI Programming Competitions

Warp, an AI programming tool, aims to make programming agents more comprehensible by emulating the concept of pair programming.

The company today launched Warp Code, a new set of features designed to give developers greater oversight of command-line-based programming agents. These features include enhanced diff tracking and clearer visibility into the agent's actions.

"Using other command-line tools feels like praying that the agent's output is something you can actually merge," said founder Zach Lloyd. With these new capabilities, he hopes to "create a tighter feedback loop for this style of agent-driven programming."

In practice, this means users can clearly see what the agent is doing and ask questions along the way. "As the agent writes code, you'll be able to see every small change it makes," Lloyd explained. "You'll have an easy way to comment on these changes and make adjustments while the agent is working."

The interface will be familiar to existing Warp users: a bottom panel for issuing direct commands to the agent, a window to view the agent's responses, and a side panel to step through each change made by the agent. You can manually edit the code, much like code-centric tools such as Cursor, or highlight specific lines to use as context for requests or questions. Impressively, Warp’s compiler automatically filters out any errors that occur during code compilation.

"This ensures you understand the code generated by the agent and that you can edit and review it," Lloyd added.

Warp's approach represents a fresh strategy in the increasingly competitive landscape of AI-powered programming. It competes not only with code-first tools like Cursor and Windsurf but also with no-code platforms such as Lovable. Major foundation model companies have also entered the fray with offerings like Anthropic's Claude Code and OpenAI's Codex—tools that Warp uses to power its own platform.

With 600,000 active users and growing, Warp remains a relatively small player in the AI coding race but is scaling quickly. According to Lloyd, the company adds $1 million in annual recurring revenue every 10 days, suggesting that many developers are willing to pay for a better way to code.