Artificial intelligence has rapidly become a headline feature in many software offerings and is increasingly integrated into their core functionality. Although companies continually unveil novel AI-driven capabilities, the most immediate impact of AI is visible in web browsers. Google has already rolled out Gemini Nano for Chrome, and Microsoft is now bringing its on-device AI to Edge.
At its Build conference, Microsoft revealed plans to expose its existing on-device AI models to web developers. This includes the Phi-4-mini model—derived from DeepSeek R1’s reasoning architecture—which closely rivals OpenAI’s o3-mini. With these new APIs, developers will be able to build fresh applications or enhance existing web apps with AI features, such as text generation, summarization, and writing assistance. Microsoft emphasizes that all processing occurs locally on the device, ensuring robust security.
These cross-platform APIs, compatible with other AI models, are currently accessible via the Edge Canary and Dev channels. Like Google’s Gemini Nano in Chrome—which already supports translation and image creation—Microsoft’s solution aims to enrich browser-based experiences.
This announcement is one among several AI-focused initiatives from Microsoft this year. Earlier, the company introduced Copilot Search in Bing during its Copilot event coinciding with its 50th anniversary, promising a transformative search experience. Additionally, Microsoft is embedding an array of AI tools into Windows through its upcoming Copilot+ PCs.
