Grand Slam Talk

Experience baseball like never before! Our AI-powered app delivers personalized MLB commentary with custom voices, styles, and insights—accessible, engaging, and built for every fan.

Comment

Grand Slam Talk: AI-Powered MLB Game Commentary

Inspiration

Baseball is a game of stories, but many fans miss out on the full experience due to language barriers, accessibility challenges, or the lack of engaging commentary. We wanted to create a platform that bridges these gaps, providing a personalized, inclusive, and dynamic way for all fans to enjoy the magic of baseball.

What It Does

Grand Slam Talk is an AI-powered application that delivers real-time and historical MLB game commentary. Users can:

  • Select historical games.
  • Customize commentator voices, styles, and languages.

How We Built It

We used a robust, cloud-native architecture to bring Grand Slam Talk to life:

  • Frontend: HTML, CSS, and JavaScript hosted on Cloud Run for fast, scalable content delivery.
  • Backend: Flask APIs deployed on Google Cloud Run for efficient, containerized processing with autoscaling.
  • AI/ML:
    • Vertex AI with Gemini 2.0 Flash for generating contextual commentary.
    • Google Text-to-Speech API for converting text into lifelike audio.
    • Google Imagen 3 for creating dynamic visuals like player stats and game highlights.
  • Data Source: MLB GUMBO API for fetching live and historical game data.

Challenges We Ran Into

  1. Real-Time Data Processing: Handling live game data with low latency was a major hurdle. We'll look into streaming with a buffer to fix this
  2. Commentary Quality: Generating natural, engaging, and accurate commentary required tuning of Gemini model prompts and would have liked deeper ssml features in Cloud Text to Speech.

Accomplishments That We're Proud Of

  • Successfully deploying a scalable backend on Google Cloud Run to handle processing.
  • Gemini 2.0 AI models to generate high-quality, dynamic commentary.
  • Making baseball accessible to a broader audience, including visually impaired users.
  • Integrating cutting-edge tools like Google Imagen 3 for real-time visuals, adding a new dimension to the fan experience.

What We Learned

This project deepened our understanding of integrating sports and AI technologies. We learned how to:

  • Process and visualize data effectively.
  • Build platforms that cater to diverse audiences.
  • Leverage cloud-native tools like and Cloud Run for scalability and performance.
  • Use AI responsibly to enhance traditional fan experiences without losing the magic of human storytelling.

What's Next for Grand Slam Talk

  • Interactive Features: Enable fans to ask questions and get real-time answers during commentary.
  • Voice Cloning: Use ethical voice cloning to recreate the iconic voices of legendary commentators.
  • Personalized Experiences: Enhance commentary based on user preferences and viewing history.
  • Expanded Sports Coverage: Extend the platform to other sports, bringing the same dynamic experience to different fanbases.
  • Monetization: Explore premium features and subscriptions for personalized content and exclusive commentary.

Grand Slam Talk is just the beginning of our journey to revolutionize how fans experience sports. By continuing to innovate, we aim to make the magic of the game accessible and engaging for everyone.

Built With

Share this project:

Updates