Skip to content

Latest commit

 

History

History
61 lines (46 loc) · 3.19 KB

README.md

File metadata and controls

61 lines (46 loc) · 3.19 KB
Raw

speechScript

Turn your voice into runnable code and demo it in your browser. Project for nwHacks 2022. Demo video can be found here.

Inspiration

What if you could just say whatever, and the code could just magically pop out? Created for nwHacks 2022, speechScript can take your lovely voice and convert it directly to code.

What it does

Converts speech to text and converts text to code; Allowing the users to use their speech to create code! Currently works with javascript, rendering the elements created by the openAI to the screen.

Technology

  • Microsoft Azure: Utilizes Microsoft Azure Speech Services to enable voice recoginition
  • OpenAI: Utilizes OpenAI's Codex engine to retrieve AI generated code from text
  • React: Designed and Implemented using the React Framework
  • Sass: Organized css files with sass, and change it back to normal css when ran

Challenges

Several challenges were enountered while building speechScript:

  • Ran into technical challenges sending data to the Microsoft Azure Speech Services API (thanks to the Microsoft team for helping us resolve them!)
  • Had to figure out the OpenAI engine landscape to figure out which API we were meant to use, and then subsequently how to properly send data to this API
  • It took a long time to figure out how to execute a string as a Javascript command in an iFrame

Accomplishments

The team made several personal accomplishments in creating speechScript:

  • Learned how to use Microsoft Azure's Speech Services API, the OpenAI Codex API, and integrate both of these with our application
  • Learned how to set up iFrames and run external commands in them
  • Learned more about CSS styling
  • Managed to successfully navigate Git while working as a team

Business viability

It can help those who might struggle using the keyboard and make coding easier. Many computer scientists and other coders alike suffer long-term hand fatigues and injuries; software like this can make their work easier and prevent injury.

Instructions

*Note* This application requires that the user have access to the Microsoft Azure Speech Services API and the OpenAI Codex engine API.

To download and set up the application:

  • Clone this git repository
  • Open a new terminal and navigate to the top folder of the project
  • Run the command npm install
  • Once this is finished, add API keys for Microsoft Azure Speech Services and OpenAI Codex services to a .env file, with the following titles:
    • REACT_APP_API_KEY = "your key here"
    • REACT_APP_SPEECH_KEY = "your key here"
    • REACT_APP_SPEECH_REGION = "your region here"
  • In the terminal, run the command npm start
  • Access the application at localhost:3000 in your browser.

To run the speechScript application:

  • Click the Record button at the bottom of the screen and begin talking
  • After a moment of silence, the application will determine what was said
  • Click Execute to generate the javascript code and run it!

Future Improvements

  • Include more programming languages for code generation
  • Automatic execution recorded speech