Anthropic introduces code review tool to manage surge in AI-generated programming

In software development, peer review is essential for catching bugs early, keeping the codebase consistent, and improving the overall quality of the final product.

The rise of “vibe coding” — using AI systems that take plain-language instructions and rapidly generate large volumes of code — has changed the way developers build software. While these tools have accelerated development, they have also introduced new bugs, security vulnerabilities, and code that users may not fully understand.

Anthropic’s response is an AI reviewer meant to catch problems before they are merged into a software codebase. The new tool, called Code Review, launched Monday inside Claude Code.

“We’ve seen a lot of growth in Claude Code, especially within the enterprise, and one of the questions that we keep getting from enterprise leaders is: Now that Claude Code is putting up a bunch of pull requests, how do I make sure that those get reviewed efficiently?” Cat Wu, Anthropic’s head of product, said.

Pull requests are the mechanism developers use to submit code changes for review before those changes are added to the software. Wu said Claude Code has dramatically increased the volume of code being produced, which in turn has increased the number of pull requests needing review and created a bottleneck for shipping software.

“Code Review is our answer to that,” Wu said.

Anthropic’s launch of Code Review — which is first arriving in research preview for Claude for Teams and Claude for Enterprise customers — comes at an important moment for the company.

On Monday, Anthropic filed two lawsuits against the Department of Defence after the agency designated the company a supply-chain risk. That dispute is likely to push Anthropic to rely even more heavily on its fast-growing enterprise business, where subscriptions have quadrupled since the start of the year. According to the company, Claude Code’s run-rate revenue has surpassed $2.5 billion since launch.

“This product is very much targeted towards our larger-scale enterprise users, so companies like Uber, Salesforce, Accenture, who already use Claude Code and now want help with the sheer amount of [pull requests] that it’s helping produce,” Wu said.

She added that engineering leads can enable Code Review by default for every engineer on a team. Once it is switched on, the tool integrates with GitHub and automatically analyses pull requests, leaving comments directly in the code to explain potential issues and recommend fixes.

Wu said the focus is on finding logical errors rather than style issues.

“This is really important because a lot of developers have seen AI automated feedback before, and they get annoyed when it’s not immediately actionable,” Wu said. “We decided we’re going to focus purely on logic errors. This way we’re catching the highest priority things to fix.”

The AI lays out its reasoning step by step, explaining what it thinks the issue is, why it could be a problem, and how it might be fixed. The system also marks issue severity with colours: red for the most serious issues, yellow for possible problems worth checking, and purple for issues linked to existing code or historical bugs.

Wu said the tool works quickly and efficiently by leveraging multiple agents running in parallel, each examining the codebase from a different angle or dimension. A final agent then combines and ranks the findings, removes duplicates, and prioritises the most important issues.

The product also provides a lighter layer of security analysis, and engineering leads can customise additional checks to align with internal best practices. Wu said Anthropic’s newer Claude Code Security product is intended for more in-depth security analysis.

Because of its multi-agent design, Wu said this can be a resource-intensive product. Like many AI services, pricing is based on token usage, and the cost depends on the complexity of the code being reviewed. Wu estimated that each review would cost around $15 to $25 on average. She said it is intended to be a premium experience, and one that becomes increasingly necessary as AI tools generate more and more code.

“[Code Review] is something that’s coming from an insane amount of market pull,” Wu said. “As engineers develop with Claude Code, they’re seeing the friction to creating a new feature [decrease], and they’re seeing a much higher demand for code review. So we’re hopeful that with this, we’ll enable enterprises to build faster than they ever could before, and with much fewer bugs than they ever had before.”