Artificial intelligence startup Hugging Face Inc. is looking to break down barriers to AI development with the launch of Gradio 5, the latest version of its open-source tool for building machine learning applications.
The latest version of Gradio is all about easing access to AI development, and includes an entirely new, AI-powered app creation tool, additional features for web development, and a stronger security focus.
Hugging Face operates a popular platform for hosting open-source machine learning projects. In some ways, it’s similar to the GitHub platform that stores the basic code for millions of enterprise and open-source applications, only it’s focused more specifically on AI models. It stores the underlying code for more than 1.3 million AI models, including large language models that power generative AI. It also hosts more than 450,000 training datasets and other technical assets.
Acting as a kind of nerve center for AI, Hugging Face has emerged as a key player in the industry and not surprisingly, it counts several major tech firms among its investors, including Nvidia Corp., which joined the company’s most recent $235 million funding round.
The company acquired Gradio back in 2021, and used that startup’s technology as the foundation of its own development studio. It has grown to become one of the key aspects of its AI platform, with more than two million monthly users and over 470,000 apps built on it to date.
With today’s launch of Gradio 5, Hugging Face has introduced an experimental “AI Playground” that makes it simple to create and preview AI apps using only natural language prompts. As Hugging Face’s Abubakar Abid, who founded Gradio, explains in an interview with VentureBeat, it’s similar to other no-code or low-code app building environments, in that you can simply enter a text prompt describing the kind of application you want to build, and it will understand that request and transform it into Gradio code.
Where it differs from other environments is it will go further, enabling users to generate an instant, web-based preview of their apps and run it in their web browser.
It’s a key innovation for Hugging Face, and given the startup’s key role in the AI ecosystem, it’s one that promises to kick open the doors to AI app development for citizen developers, while dramatically reducing the time it takes to build new apps for experienced users.
One of the main advantages of building an app in Gradio’s platform is that it’s directly integrated with Hugging Face’s AI model hub. So as soon as a model has been uploaded into its repository, developers can take it to Gradio and use it to power a new web application with just a few lines of code – and now, just a few spoken commands.
Another key area of focus with the Gradio 5 release is on bridging the gap between machine learning developers and web development, the company said. According to Abid, most machine learning developers are very proficient in using the Python programming language, but somewhat less so when it comes to actual web development, which is more reliant on a different set of programming languages and tools.
“Gradio lets developers build performant, scalable apps that follow best practices in security and accessibility, all in just a few lines of Python,” Abid told VentureBeat.
In addition, Gradio 5’s security features have been enhanced, bringing it closer to what Abid term’s “enterprise-grade”. He explained that the company hired the cybersecurity firm Trail of Bits Inc. to perform an independent audit of Gradio’s new codebase that covered all of the fixes it introduced for problems it had already identified itself.
Abid added that the latest version of Gradio is really just a work in progress, as the company is working to add plenty more features and functionalities in the coming weeks and months, including mobile support, multi-page Gradio apps, sidebars, navbars and support for new modalities around images and video.