Every data grid demo looks fast until you give it real data. This one started from the other end: a million rows in the browser, with sorting, filtering, grouping and aggregation running in a Web Worker so the page keeps responding while it works. The part that took longest was not the grid. It was the install. A worker is a separate script and its URL depends on whichever bundler you use, so most libraries hand you a README section per bundler. We compile the worker into the package and start it from a Blob URL instead, so npm install is the whole setup — in Vite, Next.js, webpack, or a plain script tag. It cost 30.1 KB (about 17%) on the core package. Measured, not estimated. KanuniLabs is two components over that engine. DataGrid does virtualized rows and columns, sorting, filtering, grouping and export. PivotGrid does drag-and-drop pivoting with a field panel your users drive themselves. Both ship for React and for plain JavaScript with no framework at all, both are typed, and both carry 12 locales with right-to-left support and six themes in light and dark. 98 of the 136 documented features need no licence key: the free tier is a real tier, with no watermark and no expiry. The paid tier adds inline editing, master-detail, charts, calculated fields, drill-down and styled Excel and PDF export, and one key covers both components. The playground on the site loads the million rows and lets you sort and filter them yourself, which is a better answer than any number in this