Microsoft is giving developers more tools to build AI-powered web apps—right inside its Edge browser. At Build 2025, the company introduced a new set of APIs that let devs plug into built-in AI models with just a few lines of code.
These new tools include access to Phi-4 Mini, a lightweight AI model released by Microsoft earlier this year. With 3.8 billion parameters, Phi-4 Mini handles math problems with ease. It was trained using both human-written and AI-generated data. Despite its smaller size, it performs well on less powerful devices, including laptops and smartphones.
This move puts Edge in direct competition with Google Chrome, which also offers built-in AI features. Microsoft aims to attract developers by focusing on privacy and speed. The new APIs work locally, keeping sensitive data on the user’s device instead of sending it to the cloud.
Edge is also gaining writing tools. These new APIs can help generate, edit, and summarize content. Soon, Microsoft plans to add a translation API to support real-time, on-page language translation.
All of these APIs are now available in Edge’s Canary and Dev channels. They are considered experimental but are built with the goal of becoming cross-platform web standards. That means they may work in other browsers too.
Microsoft emphasizes the privacy and security benefits. On-device processing makes these APIs ideal for apps handling regulated or confidential data, such as in healthcare or finance.
Alongside this update, Microsoft is rolling out PDF translation in Edge. The tool supports over 70 languages. Users can open a PDF, click the translate icon in the address bar, and get a new document in their language of choice. Canary users can try this feature now, with a full release coming next month.
With these additions, Edge is evolving beyond a browser. It’s becoming a smart platform where developers can build fast, private, AI-powered web experiences.
