OpenAI will buy Pierre, whose Trees and Diffs libraries are already in the Codex app
A post dated Wednesday says the deal needs regulatory approval, and does not give a price. Open-source projects stay open. The note points at Pando, the Git backend OpenAI says cut median push latency by nearly 5x inside its own monorepo.
Feature3 minIn-depth
James Okonkwo3 min read
- OpenAI
- Codex
- Git
- Pierre

Photo: OpenAISource page
OpenAI said on Wednesday that it will acquire Pierre Computer Company, Inc., the company behind Code Storage and a family of open-source developer tools. The post is dated 4 November. It says that anyone who has browsed files or reviewed a change in the Codex app has already used Pierre’s Trees and Diffs libraries. Closing depends on customary conditions, including regulatory approvals. Until then, the post says, the two companies stay separate and independent. It does not state a price.
The tools the post describes are for reading a change, not for writing one. Trees moves through a project’s files. Diffs shows what changed, and an Edit mode lets a reviewer correct a line in the place they spotted it. Highlights, which Pierre released recently, is a syntax highlighter the post calls extremely fast, updating as a person types or as an agent streams code. The same small team, OpenAI writes, designed the API for Code Storage, a service that lets an app store code and keep its history in Git. Pierre’s open-source projects will remain open-source, and OpenAI says it will keep contributing to them. Customers can go on using Code Storage after the close, while the company works on bringing those features into Codex.
We want Codex to connect the code a company ships with the decisions its people make, at the scale the future demands.
— OpenAI, on the Pierre acquisition, 4 November 2026
A backend the company already runs on itself
The acquisition post points at a note from Monday. Aaron Friel, Taylor Blau, and Ted Nyman, members of technical staff, wrote that OpenAI built Pando, a new backend for Git repositories, and that it now runs all development in the company’s monorepo. Engineers still use Git the way they did. The pressure they describe is Codex. More work in parallel means more branches, more pushes, more merges, and more fetches from development environments, research jobs, and continuous integration. Around the end of last year, they write, the systems under the monorepo were not keeping up: slow fetches and pushes, failed merges, and reliability that at times fell below 90 percent. In production, Pando “cut median push latency by nearly 5x,” and it “delivers nearly a petabyte of Git data per day.” It stays compatible with the Git protocol and the Git client. One limit they state directly is that a reader looking at several branches can see them at different moments, because Pando does not offer one snapshot of every reference at once.
Axios, on Wednesday, reports that the price was not disclosed and that OpenAI plans to fold Code Storage into Codex, where Pierre’s software is already in use. The story frames the deal as a challenge to Microsoft’s GitHub. It says GitHub has had reliability trouble as AI coding has grown, and that critics argue the service was built for human authors at a point when a great deal of software is written by agents. Axios names two other efforts: Cursor, which it says is now owned by SpaceX, launched a GitHub rival called Origin in August, and Thomas Dohmke, a former chief executive of GitHub, announced a competing product last month. Microsoft, Axios writes, owns nearly a quarter of OpenAI, and the two companies compete elsewhere as well. On Pando, Axios says OpenAI is exploring ways to offer the technology to customers. Monday’s post, where it describes what is running, is about OpenAI’s own monorepo, and about contributing client work upstream. It does not announce an outside product.


