Open Interpreter in 7 Minutes

--- type: transcript date: 2023-09-14 youtube_id: nL2G6Kz4BN0 --- # Transcript: Open Interpreter in 7 Minutes in this video I'm going to be showing you open interpreter which is an open source a locally running implementation of open ai's code interpreter so what open interpreter allows you to do is within a terminal instance by simply running The Interpreter command you'll be able to interact with it similar to how you would attach GPT with the code interpreter plugin turned on now the one thing that's really neat with this project is it allows you to both use open AIS llms such as GPD 3.5 turbo as well as gpt4 but it also allows you to use models like code Lama or open source models from hugging phase so to install this it's as simply as running the PIP install open interpreter command now this is assuming that you do have python installed and appropriate versions and all of that and once you've run through that you can just go ahead and simply within your terminal type interpreter now the first time that you run this you'll get a screen where you get to choose essentially two different paths so there's the open AI path and then there's the code llama path so if you want to take the open AI path you can put in your API key just like that there's also another approach which I'm going to show you in just a second here but see if you take the code llama path just click enter it will ask you to specify which model you'd like to use so let's say I want to use the 13B parameter count model you can see that the is there's the size of the model the estimated RAM and all of that now if you go through this it will actually go through and start to download the model and install everything that you need to get going with that now in this example I'm not actually going to show you that have you wait on screen and see that model download and all of that I'm just going to instead take the open AI route for things so if I run that command again interpreter and I'll show you where to go if you haven't used the openai API yet so just go to platform.openai.com API Keys you can make an account uh if it's the first time you made an open AI account you should get some credits for free simply create a secret key you can ground that and you can put it here now the one thing you will notice is as soon as you do run it for the first time so say you put in your API key you'll see that there is a tip and it says run this command and essentially this tip is so that as you exit in and out of the open interpreter here you won't have to continually put in your API keys so I definitely encourage you after you put in your API key or even just put in tests like I did just grab that line exit out export open AI underscore key paste in your key and run that command and then every time you run it like uh you'll see within your instance of the terminal if I just click interpreter I won't have that prompt to put in my API key so you see the model is set to gpd4 and I'll just quickly show you how you can change it to GPD 3.5 so if you say interpreter and all you need to do is dash dash fast so that's the flag to use GPD 3.5 Turbo now one other thing that I just want to show you is if you do want to avoid having to confirm the prompts where it asks you to run code which you'll see in just a second you can also interpreter Dash Y and you can also do this with GPT 3.5 turbo if you like so that Dash y will just circumvent it having to continue to at least ask you so the next thing that I'm going to show you is just some simple example so if I just run interpreter and I clear this so I have a empty directory as you see here so I listed out there's nothing there and I say make me uh hello World file you can see okay here's the plan I'm going to make a Hello world.txt let's start with creating a file etc etc okay so do you want to run the code yes okay now it's been created now let's move on to step two now I tried this before recording and usually it is a little bit less verbose than this example but let's just go with this approach for now so you see it's going to Echo out hello world within that file so if I just LS we see our hello world and then if I VI hello world there's our hello world so we'll just exit out of that now the I'll just exit out of this example here and you can have multiple threads within it but I'm just going to go ahead and try maybe a bit more of an involved example for you so I'm going to say I want an Express server start a new node project for me that logs out hello world from a get request so here's the plan they're going to create a directory for the node project they're going to initialize it et cetera et cetera so we're going to make a folder yes so these are the commands where if you want to circumvent and have it just execute all these commands for you you can pass in that Dash Y flag when you run interpreter so I'll go Dash y again we see okay and if I just LS out years and I'll go within the express server so we see what it's doing so there's the package Json do you want to install Express so if I would were to go in and look at the package here you can see now there's Express we'll just exit out of that okay so now it's creating the server.js do you want me to run this code yes Okay so it didn't find the command here so that's the other thing that's interesting with this project is sometimes if it makes errors it will try and self-correct itself which is pretty neat to see what I like to run this code yes so interesting approaches it's definitely a project where I wouldn't be surprised if well you start to play around with this initially that you run into some some bugs and whatnot but I found it's just a really interesting implementation especially to have an open source and be able to play around with us so I think there's going to be a lot of cool implementations of this coming out over the next several weeks and months and whatnot but very cool project so I encourage you to check it out go to the GitHub star the repo shout out to the contributors on this project excellent work really cool tool to play around with that's pretty much it for this one I just wanted to bring this to your attention but if you found this video useful please like comment share and subscribe and consider becoming a patron on page