• On the spot presentations
  • Never worry if you know your content
  • Need for reducing human effort
  • Urge to be lazy

What it does

  • Generates beautiful slides with content in real time based on the presenters' speech and context
  • Lets you dynamically add images, automatically chooses the best layout for your content
  • Eliminates the work required in formatting slides

How we built it

  • The presenters' speech is recognized using Google Cloud Speech API, then the Natural Language is parsed using to understand the context and generate data.
  • The layout changes to match the content put in each slide and images can be sourced from the web by the user by just mentioning the object name using the Microsoft Bing Search API.
  • Front-end is written in Facebook's React.js while the backend is made using node.js with webpack for bundling.

Challenges we ran into

  • Continuous background recognition of speech
  • Training of Natural Language Parser for the specific use case
  • Asynchronous calls and client/server interaction

Accomplishments that we're proud of

  • Being able to complete a fully functional prototype written with design patterns in mind
  • Having a beautiful customizable user interface that can serve both educational and professional purposes

What we learned

  • Enhancing our skillset with the state of the art tech
  • Having a logical project plan before jumping into coding really goes a long way
  • Talk is cheap, show people the code

What's next for Slides In Time

  • Ironing out bugs and getting it ready for deployment
  • Making it viiiiiral on Reddit
Share this project: