Inspiration
I always thought it would be cool to have a hologram in a museum similar to I, Robot, Foundation, or Blade Runner 2 that you could interact with a historical (or fictional) figure re-enactment.
What it does
Let's you have a conversation with a historical figure for entertainment, or more immersive learning.
How I built it
Using python, fast api, and the Nova Sonic model for speach to speach Generative AI, and the Nova Image Generator for the images I built a local dockerized application with a web front end. Using the images and the prompting the voice models I was able to make distinct personalities to reflect different historical figures to allow the user to have conversations with ai-powered reenactors to hopefully allow a more immersive, and more interesting learning experience.
Challenges I ran into
The Nova Image Generator says many of the characters go against the terms of use and it refuses to generate it. The Nova Sonic model doesn't have a Chinese, Russian, or Greek voice.
Accomplishments that I'm proud of
It is my first competition in which completed the MVP and I am proud of the result even if I know it could use a lot of work.
What I learned
How use Amazon Bedrock, OpenAI, and Nova Models and leverage them to build something pretty cool.
What's next for History2Life - An Excellent Adventure
With a little more tweaking to animate the characters and lips, and a better tuning of the persona knowledge with a knowledge base, the sourcing different voice models, I think it could be a fun education tool for students, teachers, and museums to make learning about history more fun and accessable.
Log in or sign up for Devpost to join the conversation.