If you’ve built anything with Gradio in the last two years, you’ve probably used gr.Dataframe. It’s the workhorse for leaderboards, dashboards, and any app that needs to show tabular data. But it’s also been accumulating bugs and feature requests while the team focused elsewhere. That neglect officially ended this week.
The team shipped a massive overhaul, closing over 70 issues in six weeks. The marquee feature is multi-cell selection — you can now highlight multiple cells and copy or delete values across your selection in one go. It sounds basic, but its absence forced developers into clumsy workarounds for years. Row numbers and column pinning also arrived, which means no more losing context when scrolling wide datasets. A dedicated copy button exports selected cells as comma-separated values, and a full-screen mode strips away UI chrome for focused data work.
Accessibility got a serious look too, with improved keyboard navigation that makes the component usable without a mouse. Developers gain new controls: the static_columns parameter locks specific columns against editing, while row and column selection events feed entire row data into your app’s logic. Search and filter modes — toggled via the show_search parameter — let users find or narrow data without custom code. A scroll-to-top button and a styler parameter for custom CSS round out the quality-of-life improvements.
What’s notable here is the pace. Seventy issues in six weeks suggests the team either triaged aggressively or finally prioritized a component that’s been showing its age. Either way, the update is live in the latest release. A small but telling detail: the example code in the announcement uses Irish wildlife data — red foxes, badgers, otters — a charming touch that reminds you Gradio remains a project with personality, not just another enterprise toolchain. The team is promising more demos on their X account and continues soliciting GitHub issues. If you’re maintaining a Gradio app with a dataframe, this is the release you’ve been waiting for.