Work in progress
Note : the wiki and this README are pretty poor for now (in particular for installation and Google/Amazon cloud accounts to create/configure to make this thing work) and I will take time in a near future to explain a bit more what this project is made for, and how to properly install it and make it work.
The aim is to build a full featured and easy to code/understand API to work with Snips.ai, an open source and privacy compliant voice recognition system (with a little bit of help from Google and Amazon just to make text-to-speech and speech-to-text).
$ npm install # to install all dependencies
$ npm run build # to build one time the app
$ npm run watch # to watch changes and recompile the app at each change
$ npm start # to run the app and restart it when 'watch' recompile it (nodemon)
$ node dist/index.js # to run the app (won't restart when a change occurs)May be out of date sometimes :
The assistant used in this project is available here. It will be updated over time when the project will handle more and more requests. You can also of course build your own in your snips admin console.
- A Raspberry-pi, or a Linux machine, with raspbian/debian on it
- To install snips on it (see their doc, maybe i'll write some here one day)
- An AWS account (free one) to use Polly text-to-speech
- A Google Cloud account to replace the Snips ASR which is pretty poor in vocabulary (had trouble with a lot of requests)
- Configure snips to use Google ASR instead of Snips own ASR (see links bellow)
- Clone this repo
- Start to have fun
See the wiki for more informations.