Warp launches Warp Code with tools for code review, agent profiles & integrated diff view

The AI terminal emulator Warp has released Warp Code, a suite of features that gives developers more control over command-line coding agents. It introduces a code review panel where users can see differences in real time, request modifications, and edit directly in the interface. The integrated diff view removes the need to switch to GitHub when reviewing agent-generated changes.

Projects now generate WARP.md files compatible with Agents.md, Claude.md, and Cursor rules, allowing instructions and versioning to stay consistent across platforms. Workflows are further simplified through slash commands, which enable quick project setup, rule management, and server connections, reducing the friction of moving between tasks.

To build on this, Agent Profiles let developers define models, permissions, and tasks for each agent, ensuring they operate within clear boundaries. Warp’s compiler adds another layer of support by automatically troubleshooting errors and suggesting fixes for conflicts, helping developers keep projects moving without interruptions. With its coding agent ranked #1 on Terminal-bench and #3 on SWE-bench Verified using GPT-5, Warp Code positions itself against tools like Cursor, Windsurf Editor, Claude Code, Gemini CLI or OpenAI Codex by focusing on a tighter feedback loop between developer and agent. Upcoming improvements will bring LSP support, symbol navigation, and expanded review options to build on this foundation.