ChatGPT Mirror

--- type: transcript date: 2023-03-26 youtube_id: MJicAbTD1Zs --- # Transcript: ChatGPT Mirror today I want to talk to you about a new open source web app called chat GPT mirror that's been gaining attention in the developer community in this video we'll go through how to set it up why you should potentially use it and how it can be customized to your needs to get started with Chad GPT mirror the first thing that you'll need to do is pull down the repo next we're going to grab an API key from open AI once you've obtained the key add it to a DOT EnV file in the root directory of the repo just like this after setting up the API key create a configuration file that specifies the system message and settings for chatgpt to use you might be wondering why you should considering using chatgpt mirror and I could think of a couple reasons for those not subscribing to the premium version of chatgpt chatgpt mirror can serve as a cost-effective way to access some of the more advanced models the primary advantage however is the ability to establish preset contacts for prompts allowing you to customize chat GPT according to your specific needs but before I get into that I just want to do a little bit of a breakdown of the cost structure with GPT 3.5 turbo the API rate is zero point zero zero two dollars per thousand tokens assuming an average five tokens per word you would get around 200 words for every thousand tokens so you can do a bit of a cost benefit analysis to demonstrate how chai GPT mirror could potentially be more affordable depending on your use there's also a link in the description with an interactive text box where you can get a sense on how many tokens you might need so back to customizing chatgpt Mirror by utilizing the configuration file you can direct the prompts to generate responses based on the particular topic or field that you might be interested in this means that you can essentially Prime chat GPT mirror to be customized to your specific needs whether it's for development tasks or for other purposes for instance let's say you want a web developer assistant named devgpt and you want to use chat GPT for providing coding suggestions you can do so with something like this you could set up the configuration file system message to have contacts related specific to web development such as this by doing so chat GPT mirror will generate responses tailored to your role as a web developer now you can be a lot more specific so I encourage you to play around with a system message so there's lots of information on prompt engineering and if you're not familiar with that I'd encourage you to take a look at it because with a good system message there's the potential to have a pretty tailored assistant to what you need now the benefits of this is you might not need to input as much information with each prompt if it already has that primed system message if you found this video helpful please like comment and subscribe to the channel and until the next one