[Enchancement]: Add a save command to the CLI about gpt-turbo

[maxijonson]

Sorry, just reopening temporarily until 3.1.0 can be published with this feature. I'm getting an error from NPM "Too many requests", which is quite odd considering I've published over 5 times in a single day in this library's beginnings 😅 Will try tomorrow!

[maxijonson]

Hi, I just wanted to give you a follow-up!

That's a great idea I've wanted to explore but I left it on the side due to some questions I've been asking myself on what kind of data to save, where to save it, and how to load it. Your idea of a timestamped file is a good idea. One could specify a flag (--save or GPTTURBO_SAVE environment variable) to indicate they'd like to save the conversation, optionally specifying a file name. Loading could also be done with a flag (--load or GPTTURBO_LOAD).

I'm currently working on the Web implementation of GPT Turbo (under the impl/web branch, at the time of writing), which is leading to major API changes (e.g. ditching OpenAI's Node SDK and using fetch instead, for compatibility). There's also a save functionality, which is good for laying a foundation for how I will approach this new feature for the CLI. So I'll work on this issue as soon as I'm ready to merge the Web implementation!

[maxijonson]

It was successfully published this afternoon! you should have it by installing the latest version (3.1.0) npm i -g gpt-turbo-cli. Let me know what you think! Thanks!

[jjbell150]

Thanks for adding this! It works great, so much easier than trying to copy/paste the output from a terminal