OpenAI acquires Pierre to supercharge Git and code review in Codex

The hardest part of coding with AI isn't waiting for the code to generate—it's reviewing the mountain of changes it just dumped into your repo.

OpenAI wants to fix that messy transition. The company just announced an agreement to acquire Pierre Computer Company, the team behind Code Storage and a suite of beloved developer tools, to bake modern Git workflows and deep code review features straight into Codex.

The secret engine already inside Codex

You might not know Pierre by name, but if you've opened the Codex desktop app recently, you've likely touched their work.

Pierre built Trees and Diffs, two open-source libraries that handle fast file navigation and visual change reviews. Codex already relies on both to let developers parse large codebases without breaking a sweat.

Pierre's small team is known for obsessing over interaction speed. Diffs features an Edit mode that lets a reviewer fix a broken line directly inside the diff view, rather than hunting down the file in a separate editor. The company also recently shipped Highlights, an ultra-fast syntax highlighter built to keep up as an AI agent streams tokens in real time, alongside an API called Code Storage for tracking code history over Git.

When armies of agents start branching

Why does OpenAI need Pierre right now? Basically, old-school Git wasn't designed for dozens of AI agents hacking away in parallel.

OpenAI hinted at this scaling problem earlier this week when it unveiled Pando, its in-house version control backend built to support concurrent work between human engineers and autonomous agents. Teaming up with Pierre is the logical next step on the user-facing side.

OpenAI says Pierre's open-source repositories will stay open, and existing Code Storage customers can keep using the platform while its capabilities make their way into Codex. The deal is still subject to regulatory approvals, with both outfits operating independently until closing.

Why it matters

We think this is one of the smartest tuck-in acquisitions OpenAI has made, because it attacks the real bottleneck in agentic engineering: review fatigue.

For developers, generating code has become cheap, but auditing parallel agent branches is terrifying. By bringing Pierre's responsive file trees, inline diff editing, and real-time streaming syntax highlighting into Codex, OpenAI is making the review process far less painful.

For the wider dev-tools ecosystem, this signals a massive shift. OpenAI isn't content being just the brain behind GitHub Copilot or third-party IDE extensions. With its Pando backend and Pierre's review interface, OpenAI is building a vertically integrated software development suite where the model, the branching logic, and the code review UI all live under one roof.