A modern, privacy-focused take on the classic "Guess Who?" board game, where you challenge a powerful AI that runs entirely in your browser. This project showcases the capabilities of the experimental on-device Generative AI API.

Play the game at: https://brechtdr.github.io/ai-guess-who/

This isn't your average web game. "AI Guess Who?" brings the classic deductive reasoning game to life with a unique twist: your opponent is a Gemini Nano AI model that lives on your device.

Key Features:

• On-Device AI: All AI processing happens locally in your browser. No data is ever sent to a server, ensuring 100% privacy.
• Offline Play: Once the AI model is downloaded, the game is fully playable without an internet connection.
• Custom Games: Use your device's camera to create your own set of characters and challenge the AI with familiar faces. Your custom sets are saved in the browser for easy reuse.
• Voice Input: Ask your questions hands-free using your microphone, powered by the AI's multi-modal capabilities.
• Strategic AI Opponent: The AI doesn't just answer questions; it analyzes the board to ask strategically sound questions designed to eliminate the most characters at once.

Read more in this blog post.

Because this project uses experimental web technology, a specific browser and configuration are required to run it.

1. Browser: You must use a browser that supports the on-device LanguageModel API, such as Google Chrome (you may notice better performance using Chrome Canary with the latest improvements).
2. The hosted version uses an origin trial token to enable the necessary features, but for running it locally you will need to Enable Feature Flags: Open your Chrome browser and enable the following two flags by copying and pasting the URLs into your address bar, setting them to "Enabled", and restarting the browser.

• chrome://flags/#prompt-api-for-gemini-nano (enabled)
• chrome://flags/#optimization-guide-on-device-model (place in bypass)
• chrome://flags/#enable-experimental-web-platform-features (enabled)

The first time you load the application, it will need to download the on-device AI model (a few hundred megabytes). A progress bar will be displayed. This is a one-time process.

1. Clone the repository
2. Install NPM packages:

   npm install

3. Run the development server:

   npm run dev

   The application will be available at http://localhost:5173 (or another port if 5173 is in use).

In the project directory, you can run:

• npm run dev
• Runs the app in development mode using Vite. Open your browser to the local server address to view it.
• npm run format
• Formats all project files using Prettier according to the defined style rules.