Anthropic just dropped /ultrareview (research preview) into Claude Code — a cloud-based, multi-agent code review feature that doesn't touch your local machine. Type /ultrareview in the CLI, and Anthropic spins up a team of review agents in a remote sandbox to inspect the diff between your current branch and the default branch (uncommitted changes included). You can also pass a PR number to review a GitHub pull request directly. The whole thing takes about 5 to 10 minutes, and results come back as a notification in your session.
The big difference from the local /review command is the verification mechanism: every finding is independently reproduced and confirmed by a separate agent. This isn't about nitpicking code style — it's about catching real bugs. Anthropic's docs frame the two as tools for different stages: /review for quick feedback during coding, /ultrareview for deep inspection before merging critical changes like authentication logic or data migrations.
Pricing is separate from your plan's included usage — it runs on "extra usage" billing. Pro and Max users get three free uses each (one-time, non-refreshing) before May 5th. After that, each run costs somewhere between $5 and $20, depending on the size of the change. Team and Enterprise users get no free credits. The feature requires a Claude.ai account and won't work with Amazon Bedrock, Google Cloud Vertex AI, Microsoft Foundry, or organizations that have zero-data-retention enabled.
Anthropic has been quietly testing locking Claude Code behind its $100 Max plan, so it's worth keeping an eye on how this new feature fits into the company's evolving pricing strategy.