Introducing Gradio's new Dataframe!
Hugging Face
Gradio just overhauled its Dataframe component with 70+ fixes and new features like multi-cell select and filtering. If you build ML dashboards or leaderboards, this update is a real quality-of-life win.
Based on reporting by Hugging Face — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Gradio's gr.Dataframe has quietly been one of the library's workhorses, showing up in everything from Kaggle-style leaderboards to internal dashboards. But it hadn't seen a real refresh in a long while, and the issue tracker had been piling up with complaints. The Hugging Face team spent the last six weeks knocking out more than 70 of those issues, and the result is a component that finally feels like it belongs in 2025.
The headline additions are practical rather than flashy. You can now select multiple cells at once and copy or delete across that selection, which sounds small until you've spent ten minutes clicking cell by cell in a 200-row table. Row numbers and a pinned_columns parameter let you keep key columns visible while scrolling through wide datasets, and there's a full-screen toggle plus a proper copy button for exporting values as comma-separated text. A scroll-to-top button rounds out the navigation fixes, which, again, is minor on its own but adds up when you're staring at thousands of rows.
On the interactivity side, static_columns lets developers lock specific columns from editing, useful for ID fields or labels that shouldn't change. Search and filter modes, toggled through the show_search parameter, mean users can narrow down data without writing a single line of pandas. Row and column selection events now expose the full row data, not just the clicked cell, which should make building interactive apps that react to table clicks noticeably less fiddly. Keyboard navigation and accessibility got attention too, alongside a new styler parameter for customizing appearance.
None of this is revolutionary in isolation, but stacked together it turns gr.Dataframe from a serviceable table widget into something closer to a lightweight spreadsheet UI, embeddable in a few lines of Python. The team's demo, a small table of Irish wildlife stats with pinned columns and search enabled, shows how little code is now needed to get a genuinely usable data grid. Anyone already on Gradio can grab the changes today with a simple pip upgrade, and the team says more accessibility and performance work is still coming.
My take — AI-written commentary, not fact-checked reporting
This is the kind of update that doesn't trend on social media but quietly saves thousands of developer-hours, and I respect Hugging Face for spending six weeks on unglamorous bug-fixing instead of chasing another flashy demo. Open tooling like Gradio matters precisely because it's free infrastructure for the entire open-source ML ecosystem, and a better dataframe component means better leaderboards and dashboards for everyone building on top of it, not just Hugging Face's own products.
Read more about this at: Hugging Face