Among GitHub's fastest-rising projects, one of the more consequential is not a model or an agent framework but a spreadsheet. dream-num/univer gained around 202 stars in a single day, according to the site's trending page — a modest number next to the AI projects that dominate the same list, but a steady one for a project in a less fashionable corner of software.

Univer is a TypeScript framework for building applications that look and behave like office software: spreadsheets, documents and presentations. It comes from the DreamNum team, which earlier built Luckysheet, one of the most widely used open-source web spreadsheets, and Univer is best understood as that idea rebuilt from the ground up — a plugin-based architecture, a rendering engine designed for large sheets, and the collaborative, embeddable pieces you need to put a spreadsheet inside another product.

The pitch matters more than the star count. Spreadsheets are the interface that much of the business world actually runs on, and the software that provides them is one of the last major categories with no serious open, self-hostable option: a couple of SaaS platforms and one desktop incumbent. Tools like Univer let a company embed a spreadsheet in its own product, keep the data on its own servers, and extend the grid with its own logic — the same argument that open-source databases, editors and analytics stacks have already won.

The hard part is what remains hard for every spreadsheet project: formula compatibility, at scale, across an enormous and inconsistent surface of real-world files. Univer is open about being early, and the interesting question is not whether it can out-feature a spreadsheet incumbent but whether enough products decide they want a spreadsheet they control.