Transcript for:
Guide to Installing Local AI Chatbot

I'm going to show you how to install this AI power chatbot completely privately on your computer this is running on my laptop right now and it works just like chat GPT so I could give it a prompt and it's going to give me an answer right now but the only difference is I'm not even connected to the internet this is installed on my computer and uses completely free AI models to run it so it's not going to give anything to chat GPT open AI nothing like that and I could use different types of AI models to run this chat box so I'm using llama 3 right now this is from meta AI I have one from Google I have this one called mistol I'm going to talk all about that and I want to make this completely beginner friendly and if you have no technical knowledge you could still do this in 10 minutes you're going to be able to run an AI chatbot powered by different AI models completely local without Wi-Fi on your computer and the best part you could actually press the plus sign down here once it's all set up upload documents and interact with your own documents completely privately you could turn off your Wi-Fi to do this just to make sure this data is not going anywhere I'm going to show you all of that in just 10 minutes I do also have a complete 90minut course I put together to show you all kinds of more advanced thing about this local AI chatbot like system prompting using a document library and all kinds of different things that I won't cover in this video but this video will show you everything you need to know to get to this point where you're going to get this running on your computer and choose between different models to get different AI responses let's jump in now the very first step in the process is you need to download this free software this is called olama you could find it at ama.com and just press download now it works on a Mac and it recently got released for Windows 2 and it's available for Linux I'm on a Mac here I'm going to press download the process is going to be the same on Windows and Mac from this point forward I'm going to press download for mac and if you just double click to open it I'll just press open here it just download it and I just need to move it to my applications folder so do the same thing on your Mac or your PC now it's going to bring up this popup right here and I'm going to press next and it's going to say install the command lined AMA okay so I'm going to press install here it's going to ask you to type in the passcode for your computer and then it's going to give us this code here that we need to copy okay so that's going to take us to the Second Step so AMA installing it on your computer and copying this code is the very first step now the second step is going to seem a little technical because a lot of non-technical people people have never opened this application on their computer but it's an application that is already on your windows it's already on your Mac but you probably never used it before but it's really simple let me show you how it works go to the search bar on your Windows or Mac and type in terminal and this application here is also called terminal on Windows from Windows 10 Windows 11 is called terminal open it and it's going to bring you over here again this is going to look a little technical if you've never used it before but all you really need to do is we need this code right here so I just need to copy it and I need to come over here and just paste that same code over here so it just says ol Lama space run space llama 3 now llama 3 is the open-source model from meta from Facebook basically llama 3 is a very good competitor to chat GPT but completely free to use model over here so I just have to press enter that's going to install it to my computer now it may take a few seconds here to install on your computer but that's really all needed to do now if I just type in hello what's your name and here's the answer hello I'm llama a large language model trained by the team of researchers at meta AI so already I have a large language model a chat GPT competitor installed on my computer but this is going to take place inside of terminal so technically you could use it and it works completely fine but you're going to have a lot of limitations it doesn't look that good you can't upload files to it so we're going to take it a step further now for step number three we might also want to include other models so if you go back to the AMA website and go to this models tab so we installed llama 3 that is the meta AI large language model right but there are others there is this one 53 this one is really really small so it's not going to take up much space on your computer this one is by Microsoft okay so you could install this one and as you go down the list you'll see all kinds of different competitors to chat GPT that you could install completely local without Wi-Fi and use those instead what I'm going to show you in the next step is actually going to let you pick and choose from those different models too so if you're not familiar with any of these basically llama 3 is going to be the best one by far 53 from Microsoft is going to be the smallest and it's pretty good mistol and mixol are by the same company and they're very very good too just before llama 3 came out these were the two best ones and Gemma here is from Google so those are the ones you could get so I don't recommend any more than that but there are plenty here you could explore on your own now this depends on how much hard drive space you have on your computer so the small ones take maybe 2 gigabytes 3 gigabytes but something like mix stroll is going to be a whole lot bigger so if you click on it it's actually going to tell you the file size that you need on your computer so this is 26 gigabyt of hard drive space so make sure you have that kind of space available the smaller ones like 53 that I mentioned over here this is closer to 2 gabt right very small it's 10 times smaller than the other one that I showed you so if you want to install any of these which would be the next step in the process so I'll just show you all you have to do let me take mistol here I'll click on it and if you go down here usually it has a usage section when you click on them and it's the exact same type of code so it always starts with ol l space run space the name of the large language model so all I have to do is copy this over I'm just going to close this terminal and terminate it and open a new one just to show you here from scratch you just paste that every single one is going to have the same exact formatting just these three words and press enter and it's going to go to install it here now these I have installed so you see it just takes a second for it to load but usually it will just give you some numbers and things like that here showing you that it's going through the installation process now what is your name now this one doesn't have a name I'm a AI program designed to assist with information and answer questions okay so that was our third step so we installed AMA we added a large language model in that terminal app that I showed you we did llama 3 first and then we came over here to this models page and we chose a couple of other ones you don't have to that is an optional step you could just use llama 3 for your AI chatbot now this is the next step step number four we need to go to this website called open web UI this is what's going to give us that look of chat GPT using what is already happening inside of that terminal app right the terminal app is a little bit clunky right you're not going to want to use that as your AI chat Buton now this part is probably the most technical part but let me walk you through step by step if you press download open web UI it's going to take you to GitHub this is a developer website so it's going to look a little bit technical here but let me show you exactly how this works and you only have to do this one time so the first thing we need to do is in this step is we need to install an app called doc so to install that all you have to do is go to dock. and then you could go ahead and install it from here press get started on top and then again it's going to recognize what computer you're on so if you have apple chip for example that's what I'm on I'm going to go ahead and install it and same thing as before just open that from your download folder and grab Docker and put it inside of your applications folder and now I'm going to go to my applications folder look for Docker over here there it is and I'm going to double click to open it and if you see a popup that says another application change the desktop configuration just reconfigure that over here and type in the password for your computer okay now we don't have to do anything here we need to go back to that GitHub page so that was from the open web UI where I clicked this I need open this page so Docker is what I needed to install and if I come a little bit over here it says installation with default configuration it gives us another line of code here that we need to copy and paste into that same terminal app so let's go ahead and open terminal one more time and I have a brand new terminal again I closed the last one I just need to paste that code that I got from the page behind me right here press enter and it will most likely go through some kind of installation process where it says pull complete pull complete it's kind of downloading these things that you need to run this locally with Docker so let this finish up here and then you should bring you back to the home here so all I did was copy and paste press enter nothing else here now now let's go to Docker that application that I have open here now you'll see something over here that is running okay so that triggered something in the background on your computer and this is running right now so all I have to do is right here where it says Port I just have to click this and it's going to open up this page here and we need to sign in or create an account or press sign up right here for open web UI again this is going to be completely private once we install it we just need to create ourselves an account and then it's going to bring us here where it says we're logged in okay so let's go and now I have a local install of what looks like chat GPT but powered by open source models like llama 3 on my computer so I'm going to go ahead and turn off my Wi-Fi just to show you that this does not require any web access this is going to be completely private completely installed on your computer now before I show you exactly how to get this going I should mention as far as speed goes this depends totally on your computer with chat GPT and other models like Gemini from Google those are using the power that those computers are providing through the website right this is not this is using your local computer and what it's going to use is your GPU so your graphics card is really going to change the speed of this for you so if you have a good graphics card I'm on a M3 Mac here the M3 chip Mac so it runs really really fast but that is a pretty expensive computer if you don't have a graphics card at all all it can use the CPU of your computer and it's going to be very slow so you really need a kind of a good computer here to run this locally and this is something that you'll have to test out for your computer to see if it runs slow or fast but here now all I have to do and look at the website it says Local Host 3000 this is not a website on the internet this is just on my computer I'm just using Chrome here as a browser but not connected to the web my Wi-Fi is off all I have to do first is select the model on top and this is how I showed you in the beginning how we installed llama 3 and we installed mistol so I installed some other ones the same exact way you could install all of them from that one page that I showed you in the beginning it's a really simple three words that you add to terminal so let's say I want to use llama 3 here you choose that model as your uh model here and set it as default so you could use it every single time you don't have to choose every single time so that's a really nice option and down here is where your me message could go and there you go you got the answer here from llama 3 and what's really cool about it is if you don't like that answer you could try another large language model here ask the same exact question and go ahead and do it in the same chat but this time it's going to use a different large language model to kind of compare different answers and this is going to do the same kind of thing it's going to ask you some questions and you could fill out so this part is just basic prompting that you're probably used to with chat GPT and other models okay let me go ahead and start a new chat I'm going to choose llama 3 here which is set to by default and probably the most likely reason you're watching this video is you want to interact privately with your own documents and that lets you do it right here if you press the plus sign you could upload documents to it use llama 3 to look in and summarize those documents for you or analyze them so I have this PDF here and I'm going to say give me a one paragraph summary and there we go this is actually very accurate right this is about the memory function inside of chat GPT and I wrote a little PDF book about that and it's done a good job summarizing it and I could go back and forth it also could read out loud too so it has that function built in so if it automatically does it you could go ahead and turn it off over here and you could continue the response here you could go ahead and regenerate the response you could copy the response and as I mentioned I put together a 90-minute program that shows you exactly how to use this to its full potential because this has a lot more to offer so it has system prompts and you could kind of fine-tune it and train it to be a very personalized chatbot so I teach you about system prompting it has a whole document section here where you could inut all kinds of different documents to it and it could actually connect to chat gpt2 and you could use all the different chat GPT models including vision and analyze things with vision in it now that's going to connect to open ai's API to do that but it basically lets you use chat GPT 4 GPT Vision all using API so you just pay for what you use so that's technically not going to be this private option but that is a way to get unlimited usage basically pay for as much as you need out of chat GPT if you usually hit your limit out of chat GPT when you use it on the chat GPT website this is a way to get around that using all the different models that you can't typically pick and choose in the chat GPT website you could also connect dolly3 to it and create images here all kinds of different functionalities that I cover in that course on our website skill leap. so I'll link that below you can start with a free trial and that also gives you access to all our courses so we have over 20 courses you could pick and choose we just also rolled out a new assessment a simple multi-question assessment that gives you very recommended courses so you know exactly based on your skill set where you should start and based on your profession so that is going to be linked in the description below but that's how you run a local AI chatbot on your computer without Wi-Fi to privately interact with your own documents and have the power of AI without using the internet I hope you found this useful I'll see you next time