Skip to content
A web-based speech-to-code editor for humans.
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.
.ebextensions Waaargh Jan 5, 2019
src Use arrow function for class methods May 30, 2019
.npmrc Once more into the breach Jan 3, 2019
package-lock.json Update dependencies. May 27, 2019
yarn.lock Initial commit from Create React App Dec 17, 2018

Welcome to CodeSpeak!

CodeSpeak is a web-based code editor that lets users code with the power of their voice. It's currently in an embryonic stage, but my intent is to grow the product into a mature, multi-language tool.

I'm a legally blind software engineer, and although CodeSpeak isn't ready for prime-time yet, my hope is that tools like this will aid developers from all walks of life.

Getting Started

The live version of CodeSpeak can be found at - or you can clone this repository and run it locally.

CodeSpeak uses the Speech Recognition API, which currently works best in Google Chrome. Most other browsers don't support it at all, or support it only partially, but its adoption is growing, so CodeSpeak should work on other browsers soon.

To set up locally, simply clone this repository and run npm install from its root directory.

CodeSpeak also works best using HTTPS. This is enforced on, but if you decide to run it locally you may need to use the following command to make things work right:

  • Windows (cmd.exe): set HTTPS=true&&npm start
  • Windows (PowerShell): ($env:HTTPS = "true") -and (npm start)
  • Mac/Linux: HTTPS=true npm start

And, running locally, you'll still need to proceed past the usual "This site is unsafe" warnings.


As I said, CodeSpeak is in its embryonic stage. Right now it only supports Python, and only has a limited number of commands:

  • "Hello world.": enters the obligatory print('Hello world') command.
  • "Create function [function name]": Creates a def function_name(): stub, with no arguments yet.
  • "Create variable [variable name]": Creates a variable with the specified name.
  • "Set value [variable value]"": Sets the value of the created variable.
  • "Goodbye world.": clears the editor window.

More commands and functionality coming soon, promise!


CodeSpeak is built on the following technologies:

  • React: This is overkill right now for what CodeSpeak is/does, but I hope it will give me room to scale the app as it grows.
  • React Ace: A react implementation of the Ace editor, created by Cloud9. This is the WYSIWYG editor interface.
  • annyang: This is the JavaScript speech engine that powers CodeSpeak.


Everything! But here's a short list:

  • More commands: Need to add enough Python commands to at least allow a user to write a very simple program entirely by voice.
  • UI/UX I'm not great at UI, but I know it needs some work.
  • Save/share functionality: Save code snippet to file, share as gist, etc.

More ambitious goals:

  • Other languages: I'd like to have this working in other languages besides Python.
  • Integrated interpreter: I'd like users to be able to actually run their code snippets in the browser.
  • Text-to-speech: The flip side of speech-to-text, I'd like the editor to be able to give verbal output.


Yes please! This is my first open-source project, and, I'll be honest, I don't quite know what I'm doing. So help on any level would be greatly appreciated!

You can’t perform that action at this time.