Microsoft Edge Integrates On-Device AI APIs for Web Developers

On May 19, 2025, Microsoft rolled out experimental on-device AI APIs for its Edge browser in the Canary and Dev channels. These tools allow developers to run text-based AI features directly inside the browser using local resources.

The key addition is the Prompt API, powered by Phi-4-mini, Microsoft’s lightweight 3.8B parameter language model. Developers can now build features like text generation, rewriting, and analysis with a few lines of JavaScript, no external servers or API calls needed.

Also read: Microsoft Edge Is Finally Testing New Sorting Options For Favorites

Microsoft also introduced Writing Assistance APIs, which include Summarizer, Rewriter, and Writer. These tools support tasks like condensing long articles, refining user-generated content, and composing drafts. Everything runs locally, which boosts performance and sidesteps privacy concerns tied to cloud processing.

These APIs aim to give developers easier access to fast, private AI capabilities inside web apps. They work on Edge across Windows and macOS and reflect Microsoft’s push to set future browser standards through the W3C.

For now, these APIs remain experimental, and Microsoft plans to gather feedback before broader rollout. Developers can start testing today in Edge Canary and Dev builds. This move marks a shift toward integrating AI into browsers without needing roundtrip calls to the cloud.