Qodo 2.1 boosts code review accuracy by 11%

The code review startup Qodo has launched Qodo 2.1, introducing an intelligent rules system to tackle the 'amnesia' issue of AI coding tools. This system improves code review accuracy by 11%, eliminating the need for temporary solutions like saving states in text files.

The new system replaces static rule files with a smart governance layer that automatically generates rules from coding patterns and past review decisions. It continuously maintains rule health and reinforces standards in each code review, delivering tangible improvements.

According to Qodo's CEO, Itamar Friedman, the system marks a fundamental shift in AI code review tools, making them proactive rather than reactive. This approach has led to a 11% improvement in precision and recall, identifying 580 defects across 100 real PRs.

For companies, Qodo 2.1 offers flexible deployment options, including self-service SaaS and on-premises hosting, addressing data governance concerns. The pricing model remains seat-based, with plans ranging from free to customized for large enterprises.

Early users, such as HR tech company Hibob, report positive results with the system, noting increased consistency and improvements in code review quality. Founded in 2018, Qodo has raised $50 million from prominent investors.