Inspiration
The inspiration came from my sister. Often times due to her dyslexia she struggles to read electronic works without Mac accessibility features and as a result I wanted to try to make an editor that had audio recital support within it.
What it does
This in short is a text editor that enables writing and recital of your text. It further allows for nodal editing where you can save a version of your file and roll back and forward through it's history as you like.
How we built it
This was done in one-shot within the "bolt.new" prompt window
Challenges we ran into
This required a great deal of prompt engineering as the AI continuously got a few innocuous things incorrect multiple times. It resulted in the use of multiple simultaneous instances of "bolt.new" AIs running to isolate what was the most effective way to convey the desired outcome.
Accomplishments that we're proud of
We're really happy that the rollback and forward is so seamless and how well the audio is integrated into the overall experience. The specificity with which you can play back audio is wonderful.
What we learned
There was a great deal of learning with regard to the specificity of detailing necessary to get the AI to perform exactly as I wanted. By the end of it I felt as if I was writing a specifications sheet for engineering design. Further I learned quite a bit about already existing web speech tools in the JS API.
What's next for audiodoc.online
This is more of a first pass at the issue but a lot more will hopefully be done, from a stand alone desktop application to further refinement of the tooling, hotkeys, and UI. Further for the stand alone application a permanent file store rather than in memory store is really necessary to make this a fully fledged text editor.
Built With
- bolt.new
- elevenlabs
- javascript
- netify
- react
Log in or sign up for Devpost to join the conversation.