Share

DEC 11, 2024

Volley Prototypes an Audio-First Game with Gemini 2.0 Flash and the Multimodal Live API

Max Child

Volley

Vishal Dharmadhikari

Product Solutions Engineer

Volley showcase hero

The Gemini API is empowering developers to build the next generation of immersive experiences, and Volley is leading the charge in the world of voice-controlled AI games. Known for their hit games like Jeopardy! and Song Quiz, Volley is leveraging the cutting-edge capabilities of Gemini 2.0 Flash, currently in experimental preview, to prototype a new audio-first twist on the classic game, 20 Questions.

Volley has captivated millions with engaging voice-powered games across smart TVs, Amazon Alexa, Google Assistant, and mobile platforms. Now, they're setting their sights on a new frontier: transforming casual gaming with the power of generative AI’s live, multimodal capabilities — starting with 20 Questions.

Gemini 2.0 Flash: The Key to Immersive Voice Gameplay

Volley’s new 20 Questions prototype uses key features of Gemini 2.0 Flash to create a truly unique experience that goes beyond the game’s current AI capabilities. While the classic game features dynamic content generated on the fly, Gemini 2.0 Flash transforms it into something extraordinary - bringing lightning-fast responses, emotive new personalities, and a conversation flow that feels remarkably human.

A prototype of Volley’s 20 Questions: One of the first games powered by Gemini 2.0 Flash and the Multimodal Live API (sequence shortened)

Here’s how Gemini 2.0 Flash helps:

  • Low-Latency Dynamic Question and Response Generation: Gemini 2.0 Flash’s native audio output and the low-latency interactions unlocked by the new Multimodal Live API enable dynamic conversations with an AI Riddlemaster. The prototype enables a natural, back-and-forth conversation with the Riddlemaster: asking questions, getting hints, and more. The combination of model intelligence and long-context memory ensures personalized experiences that evolve in real-time, based on player interactions. Sub-second latency enables a truly natural human-like conversation.

  • Voice Activity Detection: The API’s built-in ability to allow for natural voice interruptions to the model’s responses enables fluid, accessible conversations, without the need of visual or haptic input.

The Future of Voice-First Gaming: Powered by Gemini

Volley envisions a future where voice AI is at the crux of gameplay, creating accessible and immersive experiences for everyone. The company’s commitment to AI innovation positions them at the forefront of this exciting new frontier. As Co-Founder and CEO Max Child elaborates, “LLMs and voice recognition technology are transforming games, breathing life into play through dynamic, interactive experiences. They enable players to immerse themselves in lively, engaging adventures where their voices truly drive the story."

Volley’s newest 20 Questions game, powered by Gemini 2.0 Flash’s Multimodal Live API, is still in the prototyping phase, but stay tuned for more information soon.

Getting Started with the Gemini API: Build Your Own Interactive Worlds

Volley's work with Gemini 2.0 Flash and the Multimodal Live API showcases the exciting possibilities of AI in gaming, particularly the potential for dynamic gameplay, lifelike characters, and natural-sounding conversations. As a game developer, you can harness the power of the Gemini API to create similarly immersive and innovative experiences.

Explore the Gemini API documentation and discover how its capabilities can empower you to build the next generation of engaging and inclusive games.