—

✨ Build Challenge ✨

Overview

This application is written using Nodejs and Express. It enables a user to have their speech transcribed using their microphone onto a rich text editor. They now have a choice between using spoken or written text and can interchange the two depending on how they would like their final product to be formatted.

Submission Category:

Accessibility Advocates

Link to Code on GitHub

• Link

  morehwachege / potential-memory

  Bonga is Swahili slang for speak/talk. Bonga Editor is a JavaScript application that is primarily meant to turn speech into text in real time hence it can be used to write an article or produce transcripts while a podcast is running in the background.

  Bonga Editor

  Bonga is Swahili slang for speak/talk Bonga Editor is a JavaScript application that is primarily meant to turn speech into text in real time hence it can be used to write an article or produce transcripts while a podcast is running in the background.

  

  Installation

  ---

  A little intro about the installation.

  Use the package manager npm to install Bonga.

  $ git clone https://github.com/morehwachege/potential-memory.git
  $ cd ../path/to/the/folder
  $ npm install
  $ npm start

  Usage

  Start the server at port 3100

  node index.js

  Contributions & Collaborations

  Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

  Please make sure to update tests as appropriate.

  Guidelines

  • Be informative. Format your pull requests nicely. Include screenshots if applicable.
  • Be a good citizen. Try your best to adhere to the established styles of the project. This doesn't mean that you shouldn't…

  View on GitHub

Live Screen Editor Platform

—

Introduction

Deepgram's Speech to text technology presents endless possibilities. The scope of products you can produce using speech are endless.

I decided to enter this hackathon just to have a taste of automation and making work easier since tech mostly focuses on that.
Speech to text == commands is a logic that has changed the world.

My Deepgram Use-Case

A while back when I started out on writing articles. My goal was to document everything that I had learnt thus challenging me not to lag in my journey.

My problem was that sometimes after a long coding sessions and debugging my hands got too tired to type hence often opted to procrastinate on both the article and my personal standups till I had more time to spare.

Having that I can't cut out coding time I skipped most sessions.
I was so excited to start this project because it closely relates to a problem that I have faced and providing a solution to that is intriguing for me.
During this process I have encountered so many different technologies, most of which are new to me including Deepgram.

Challenges

I wanted to learn Nodejs and Express which I dedicated to this project. Integrating Deepgram was a breeze.

However, the editor has overall been the greatest challenge yet.
I went over very many rich text editors but I couldn't seem to understand how to get the transcribed text into the editor's inner frame.

Editor APIs considered in development:

• CKEditor
• TinyMCE
• Joomla

None of these worked for me at the time so I ventured to find a way to create my own. I now prefer this version since I can build on it as I go. Languages used for the editor are html, css and javascript.

Deep Dive and Details

This application works for content creators or anyone who finds it easier to speak than type at any moment.

Using the wysiwyg editor you can get your speech transcribed and formatted and only copy when you are satisfied with the results.

For the most part Content Creators would benefit the most from this product.

• Podcasts and radio sessions can instantly be turned into blogs.

• Article writers can speed up their production since speech is faster than typing.

• Users may use the product to transcribe meetings with plausible formatting to produce easily readable material through Voice Activity Detection. This is just like real time minute recording, only a lot less effort is involved.

• Since you don't really need hands on this one, people with disabilities can now make a few more dollars on online content. It gives them a 'voice' in the industry.

• Learning platforms - Teachers can publish academic papers off of material they teach. Learners can have a record of club proceedings.

I have learnt a lot in terms of speech-to-text technology and speech recognition AI.

Some of the problems such as the mic picking up ambient noise and surrounding sounds may sometimes distort a transcript. Deepgram has given me an in depth introduction and my research into natural language processing has been eye opening.

Some of the features offered help to separate between age groups. A younger age group will have enforced profanity filters whilst an adult age group will have a choice between activating the filters or not.

Conclusion

Participating in this Hackathon has been an amazing experience overall. Learning by doing increases retention rate and I appreciate this platform for such a chance.

—