Anthropic Launches Claude Code Ultrareview to Hunt Bugs With Cloud Agent Fleets

Apr 26, 2026 · Updated May 4, 2026

Anthropic launched /ultrareview in research preview for Claude Code, its terminal-based agentic coding tool. The command triggers a fleet of bug-hunting agents in a cloud sandbox to perform deep inspections. Unlike standard reviews, this multi-agent swarm (a coordinated group of specialized AI agents) operates remotely to identify complex errors.

This update follows multi-agent code reviews and extends Claude Code session recaps designed to manage complex agentic workflows. Moving the workload to the cloud allows parallel agents to go beyond simple pattern matching to actively stress-test critical infrastructure using increased inference-time compute (allocating more processing power during the response phase).

Access the feature by running claude update to reach the latest version, which also adds to Claude Code session handoffs between mobile and terminal. Pro and Max tier users get three free reviews through May 5. It is recommended for critical changes like data migrations where human review might miss edge cases.

View the full update on code.claude.com

ClaudeDevs

@ClaudeDevsApr 22

New in Claude Code: /ultrareview (research preview) runs a fleet of bug-hunting agents in the cloud. Findings land in the CLI or Desktop automatically. Run it before merging critical changes—auth, data migrations, etc. Pro and Max users get 3 free reviews through 5/5. https://t.co/OEzhNHcZdE

1.2k17k

View on X

Still wondering? A few quick answers below.

Claude Code ultrareview is a research preview feature that dispatches a fleet of autonomous bug-hunting agents to inspect your code in a cloud sandbox. It is specifically designed to find complex logic errors in critical files, with all findings automatically synced back to your local command line interface or desktop application for immediate review.

To use the new feature, you must first run the claude update command to install the latest version of the Claude Code tool. Once you have updated, you can initiate the cloud-based agent fleet by entering the ultrareview slash command. This process uploads your code to a remote environment for deep multi-agent analysis and verification.

Anthropic has made the ultrareview feature available to users on the Pro and Max subscription plans. These users are granted three free cloud-based reviews as part of the research preview period, which lasts until May 5, 2026. This allows developers to test the high-intensity bug-hunting capabilities on their most important pull requests at no extra cost.

Anthropic recommends running an ultrareview before merging high-impact changes like authentication logic, security protocols, or data migrations. Because the command uses a fleet of agents in the cloud rather than a single local model, it provides a more thorough inspection of your code to catch subtle bugs and regressions that might be missed during standard development cycles.