Warp has unveiled Warp Code, a new feature set designed to make AI coding agents more transparent and collaborative, closer to a pair-programming experience. The update allows developers to track every change an agent makes, review diffs, and edit code in real time, creating a stronger feedback loop.
The interface maintains Warp’s familiar layout, featuring a command input field, an agent response window, and a step-by-step side panel displaying edits. Users can adjust code directly, highlight lines for context, and rely on Warp’s compiler to automatically resolve compile errors. Founder Zach Lloyd said the goal is to ensure developers can understand and refine AI output instead of blindly relying on it.
Warp enters a competitive market against tools like Cursor, Windsurf, Lovable, and command-line agents from Anthropic and OpenAI. While Warp uses models from those firms, it differentiates by focusing on transparency and collaboration. The startup is growing rapidly, reporting 600,000 active users and $1 million in new ARR every 10 days.
