Prompt API Enables On-Device AI Interactions Through Gemini Nano in Chrome
Revolutionary Browser-Based AI Tool Brings Natural Language Processing Directly to Users
Google has introduced the Prompt API, allowing developers to send natural language requests to Gemini Nano directly within Chrome browsers. The API, first available through origin trials in Chrome 138 and expanded in Chrome 148 with sampling parameters, represents a significant shift toward on-device artificial intelligence capabilities.
Practical Applications Across Industries
The Prompt API opens numerous possibilities for developers seeking to integrate AI functionality into web applications and Chrome Extensions. Key use cases include:
- AI-powered search: Answer user questions based on webpage content
- Personalized news feeds: Dynamically classify articles by category and enable user-driven filtering
- Custom content filters: Automatically blur or hide content based on user-defined topics
- Calendar event creation: Extract event details from web pages for streamlined calendar entry
- Seamless contact extraction: Pull contact information from websites to simplify business communications
System Requirements for Implementation
Developers and end users must meet specific hardware and software requirements to utilize the Prompt API. The technology functions exclusively on desktop platforms including Windows 10 and 11, macOS 13 and later, Linux, and ChromeOS version 16389.0.0 and onwards on Chromebook Plus devices. Mobile versions of Chrome, along with non-Chromebook Plus ChromeOS devices, currently lack support.
Technical specifications demand at least 22 gigabytes of free storage space on the drive containing the Chrome profile. For processing power, users may choose between GPU acceleration, requiring more than 4 gigabytes of video RAM, or CPU-based processing, requiring 16 gigabytes of system RAM and at least four processor cores. Audio input functionality specifically requires GPU support. Network connectivity must be unlimited or unmetered.
API Development and Documentation
The Prompt API documentation, authored by Thomas Steiner and Alexandra Klepper, was published on May 20, 2025, with last updates on September 21, 2025. The technical documentation covers session creation, multimodal input handling supporting text, audio, and image formats, and advanced features including context management and structured output through JSON Schema implementation.
Developers can test the API locally by enabling specific Chrome flags: chrome://flags/#optimization-guide-on-device-model and chrome://flags/#prompt-api-for-gemini-nano-multimodal-input. Multiple demonstration applications are available, including the Prompt API playground, a Mediarecorder Audio Prompt demonstration, and a Canvas Image Prompt demonstration for web users, alongside a Chrome Extension demo with source code on GitHub.
Community Engagement and Feedback Mechanisms
Google actively encourages developer participation through multiple feedback channels. Contributors can submit bug reports and feature requests for Chrome implementation, comment on existing issues or open new ones in the Prompt API GitHub repository, and join the early preview program to shape future API development.