Qodo integrates organizational context into Codex for enhanced AI code review

Qodo is developing AI tools for software quality and code review, integrating organizational knowledge directly into coding agents. The company utilizes OpenAI models, including GPT-6 Astra, to enhance its Codex plugin, allowing it to review code while accounting for complex organizational standards and codebase rules across potentially thousands of repositories. Co-founder and CEO Itamar Friedman noted that this capability allows Codex to consider dependencies across various Git providers and services connected via Kafka, helping to prevent critical P0 bugs by viewing the software holistically. Before code generation, the plugin retrieves relevant rules and pull request history, and Qodo's Agentic Toolbox reviews the changes within the same session before a developer submits a pull request. Nnenna Ndukwe, Qodo’s AI Developer Relations Lead, observes developers automating the handoff between Codex's coding and Qodo's independent review. Furthermore, Itamar mentioned an opportunity to leverage OpenAI’s Decisions API to streamline the review process, potentially selecting from hundreds of rules to speed up and lower the cost of reasoning. This approach aims to create a software factory where human expertise, AI capabilities, and traditional engineering practices converge, allowing developers to focus judgment on unresolved issues.