Using artificial intelligence locally on your computer is becoming easier than ever. Instead of relying on cloud services, you can now run many AI models directly in your web browser. This means performing tasks like summarizing text, translating languages, or generating new content without needing an internet connection or powerful hardware. Google Chrome and Microsoft Edge are leading the way by offering experimental APIs that make this possible right inside your browser window.

How Browser-Based AI Is Changing the Game

Traditionally, AI models required significant server resources and complex setups. You had to install third-party applications or manage infrastructure to run them locally. Now, browsers like Chrome and Edge are supporting APIs that allow these models to work directly on your device. This shift reduces the need for bulky software and makes AI more accessible to everyday users. With these APIs, you can execute inference jobs—like translating text or summarizing documents—on demand, all within your browser.

Both browsers are built on the Chromium project, which provides a common foundation. However, each supports different sets of AI features and models. For example, Chrome currently uses the Gemini Nano model, while Edge runs the Phi-4-mini model. These models handle tasks like summarization, language detection, and translation. Performance may vary between browsers, with some tasks running faster in one than the other, but both demonstrate the potential of local AI processing.

Key AI APIs Available in Chrome and Edge

The APIs available in these browsers are designed to perform various AI functions. In Chrome, users can access the Translator API for language translation, the Language Detector API to identify text language, and the Summarizer API to condense lengthy content into concise summaries. These APIs are ready to use and require no special setup. Edge supports most of these features as well, except for the Language Detector, which is planned for future release.

In addition, both browsers offer experimental APIs that are still under development. These include the Writer API for generating text from prompts, the Rewriter API for editing existing content, the Prompt API for making natural language requests, and the Proofreader API for checking spelling and grammar. These tools are provided on an opt-in basis and are expected to become more stable over time. The goal is to eventually standardize these APIs for broader web use, but for now, they remain browser-specific features.

Using the Summarizer API as an Example

One of the easiest APIs to try out is the Summarizer API. It can take long articles or documents and create brief summaries or bullet points. To get started, you need to create a simple web page that calls this API. The process involves writing some basic HTML and JavaScript to send text to the API and display the summarized output. This example serves as a good template for exploring other APIs like translation or text generation.

The key is to load your text into the webpage, then invoke the Summarizer API through JavaScript. The API runs models locally, so it doesn’t depend on internet speed or external servers. While performance varies between browsers and models, the experience shows how powerful and flexible local AI can be. Developers and hobbyists alike can experiment with this setup to bring AI features directly into their websites or tools.

Overall, the ability to run AI models locally within a browser opens up new possibilities for privacy, speed, and ease of use. As these APIs mature and become standardized, they could transform how web applications integrate AI, making advanced features accessible to anyone with a modern browser and a bit of coding know-how.

Inspired by

• https://www.infoworld.com/article/4154522/tap-into-the-ai-apis-of-google-chrome-and-microsoft-edge.html

Sources

• comfy.org
• lmstudio.ai
• chrome.com
• chrome.com
• chrome.com