Qodo Adds Codebase Context to Codex Reviews

Qodo, a startup that builds AI tools for code review and software quality, has added organizational knowledge and codebase standards to Codex through a plugin in the Codex marketplace. The company says the integration helps coding agents understand the code they change and catch P0 bugs across repositories. Qodo uses OpenAI models, including GPT-6 Astra, to gather engineering knowledge, review code, and monitor software quality. Co-founder and CEO Itamar Friedman says the company wants both frontier reasoning models and fast models.

Friedman says some customer environments span 10,000 or more repositories, where a seemingly correct local change can break a downstream service. Qodo’s plugin gives Codex codebase knowledge, organizational rules, and independent review through Qodo’s Agentic Toolbox. It can account for dependencies across repositories hosted by different Git providers or services connected through Kafka, helping Codex assess the software across the whole organization.

Before Codex writes code, the plugin can retrieve relevant codebase rules, pull request history, and patterns. After the changes, Qodo reviews them and returns findings to Codex in the same session, before the developer opens a pull request. Qodo’s AI Developer Relations Lead Nnenna Ndukwe says developers use this workflow with GPT-6 models in Codex to build features and address technical debt, automating the handoff between coding and independent review.

Friedman describes coding and review agents as complementary roles, each needing its own setup and context and sometimes involving multiple agents. He envisions agents discussing findings and surfacing disagreements for human review, with developers acting as planners and reviewers who can inspect unresolved issues, agent discussions, context, and prior developer decisions. Ndukwe says she uses GPT-6 Astra’s Extra High and Ultra effort settings in automated development and knowledge work. She schedules Codex to report weekly accomplishments and areas to improve, research social-platform conversations about agentic engineering, suggest developer-content ideas, and analyze her code changes to recommend agent skills to create or refine.

Friedman says Qodo could use OpenAI’s recently launched Decisions API to select applicable rules and agent skills at each development stage instead of relying on language models for every choice. For example, it could select which of 200 rules need checking before a language model reasons through the relevant ones; he says this could make selection faster and cheaper while improving review quality. Qodo describes its software-factory concept as a customer-owned system combining coding and review agents, company knowledge, traditional software engineering and scripting, and human expertise to resolve remaining questions.

Qodo, a startup that builds AI tools for code review and software quality, has added organizational knowledge and codebase standards to Codex through a plugin in the Codex marketplace. The company says the integration helps coding agents understand the code they change and catch P0 bugs across repositories. Qodo uses OpenAI models, including GPT-6 Astra, to gather engineering knowledge, review code, and monitor software quality. Co-founder and CEO Itamar Friedman says the company wants both frontier reasoning models and fast models.

Friedman says some customer environments span 10,000 or more repositories, where a seemingly correct local change can break a downstream service. Qodo’s plugin gives Codex codebase knowledge, organizational rules, and independent review through Qodo’s Agentic Toolbox. It can account for dependencies across repositories hosted by different Git providers or services connected through Kafka, helping Codex assess the software across the whole organization.

Before Codex writes code, the plugin can retrieve relevant codebase rules, pull request history, and patterns. After the changes, Qodo reviews them and returns findings to Codex in the same session, before the developer opens a pull request. Qodo’s AI Developer Relations Lead Nnenna Ndukwe says developers use this workflow with GPT-6 models in Codex to build features and address technical debt, automating the handoff between coding and independent review.

Friedman describes coding and review agents as complementary roles, each needing its own setup and context and sometimes involving multiple agents. He envisions agents discussing findings and surfacing disagreements for human review, with developers acting as planners and reviewers who can inspect unresolved issues, agent discussions, context, and prior developer decisions. Ndukwe says she uses GPT-6 Astra’s Extra High and Ultra effort settings in automated development and knowledge work. She schedules Codex to report weekly accomplishments and areas to improve, research social-platform conversations about agentic engineering, suggest developer-content ideas, and analyze her code changes to recommend agent skills to create or refine.

Friedman says Qodo could use OpenAI’s recently launched Decisions API to select applicable rules and agent skills at each development stage instead of relying on language models for every choice. For example, it could select which of 200 rules need checking before a language model reasons through the relevant ones; he says this could make selection faster and cheaper while improving review quality. Qodo describes its software-factory concept as a customer-owned system combining coding and review agents, company knowledge, traditional software engineering and scripting, and human expertise to resolve remaining questions.