Developer Tools

gradio@6.10.0

The popular AI app framework now lets developers run standalone servers and profile backend performance.

Deep Dive

The Gradio team has launched version 6.10.0 of their widely-used open-source library, introducing several major features aimed at professional developers. The headline addition is a new `gradio.Server` mode, which allows a Gradio application to run as a standalone, long-lived server process, moving beyond its traditional use as a demo or prototyping tool. This is complemented by new backend profiling capabilities, enabling developers to trace server performance and run benchmark scripts to identify bottlenecks. The update also introduces a set of core CSS themes and a `custom_css` parameter for the theme class, giving teams more control over UI styling without deep front-end work.

Other significant improvements focus on developer experience and stability. New Gradio Prediction CLI commands allow models to be invoked directly from the command line, facilitating testing and integration into larger pipelines. The `gr.HTML` component gains a `watch` function for JavaScript triggers and support for injecting custom libraries into the page head. On the fixes front, the release addresses a critical `DuplicateBlockError` in multi-page apps, resolves an issue where edited messages disappeared in the ChatInterface, and improves the reliability of custom components in Server-Side Rendering (SSR) mode. These changes collectively make Gradio more robust for production use cases.

Key Points
  • New `gradio.Server` mode enables standalone, long-lived deployment of AI applications.
  • Backend profiling tools and new CLI commands aid in performance benchmarking and testing.
  • Adds core themes, custom CSS parameters, and fixes for stability in multi-page and chat apps.

Why It Matters

This release transforms Gradio from a demo builder into a more capable framework for deploying and monitoring production-grade AI applications.