How To Install SuperAGI

How To Install SuperAGI - Multiple Agents, GUI, Tools

How To Install SuperAGI - Multiple Agents, GUI, Tools, and more

https://www.youtube.com/@matthew_berman

Transcript:

This is super AGI it's essentially Auto GPT on steroids it can use tools it can run multiple agents in parallel it has a graphical user interface it's super easy to install and they're adding new functionality every single day welcome back in today's video I'm going to show you how to install super AGI and also how to use it has a lot of functionality I'm going to show you around some of it and then I encourage you to play around with it let's go so this is the GitHub repo for super AGI it's by Transformer Optimus and it's currently on GitHub trending it has a lot of functionality let's review some of it really quick and then I'm going to get into the installation guide so first obviously you can provision spawn and deploy autonomous AI agents a couple things that it does better than Auto GPT is the fact that you can extend it with a bunch of different tools right out of the box it also has a graphical user interface and you can run agents concurrently here are some of the tools that it has so slack email Google search GitHub, Zapier, Instagram and it's all wrapped up in a Docker image so it's super easy to install let me show you how to do that the first thing you're going to need to do is get Docker so go to docker.com download Docker install it and then have it up and running this is what it looks like and you could just keep that on in the background next we're going to open a visual studio code then we're going to click this little button in the top right that says toggle panel that'll open up our terminal within Visual Studio code next on the repo page we're going to click this little green code button and then we're going to copy the URL once the terminal opens up we're going to type CD change directory and change to our desktop from there we're going to type git clone and then paste that URL that we just copied hit enter and that'll clone it to our desktop then we're going to come up to this Explorer icon we're going to click it we're going to click open folder and then we're going to select Super AGI and click open then again in the top right we're going to toggle panel so that'll open up our terminal and while that's opening the next thing we're going to do is right click on config underscore template.yaml and we're going to rename it and we're just going to rename it fig.yaml and that's where we're going to store all our environment variables so this is where you're going to put your pine cone key your OpenAi key your Google key everything anything else that you want to use this is where you put it so for example if you want to use email you can do it from here today we're just going to be using open AI Pinecone and Google search the next thing you're going to do is create an API key with open AI so if you don't already have an account sign up for OpenAi then navigate to platform.OpenAi.com account slash api-keys we're going to click create new key right here we're going to name it super2 because I already have a super then I'm just going to say create secret key I'm going to copy it and don't worry I'm going to revoke all of these Keys before publishing this video then I'm going to come back to the repo specifically the config.yaml file.

I'm going to find the open AI key right there and I'm going to paste it and save the next thing we need is a pine cone API key so if you're using a free Pinecone account you can only have one index at a time I've already created one so I'm going to go ahead and delete it and then on the left over here we're going to click API keys I'm going to click create new API key and then enter the key name I'm going to call it super2 again create key I'll copy the key value we're going to switch back to visual studio code and then at the very top it says your Pinecone API key we're going to paste over that but we also need the Pinecone environment so if I switch back this is the Pinecone environment right there I'm going to copy that as well and I'm going to paste it and save the next thing we're going to do is get the Google API key so you're going to come to Google cloud services right here under API and services click it now I already have one called super AGI so I'm just going to use that again but if you don't you're going to click this and create a new project then up here you can search for different resources so you're going to type custom search API and then under Marketplace it says custom search API you're going to go ahead and click that manage create credentials and then you're just going to say custom search API right there we're going to use the application data option and no we're not using any of these then you click next then from here we're going to type super AGI it's going to create the service account ID automatically or can generate it and then you can have a description but you don't need it create and continue select a role this is optional we're going to skip over that and then Grant users access to this account and we don't need to do that either so I click done then on the left side we're going to click credentials at the top we're going to click create credentials and we need an API key it's creating one now I'm going to copy that close it we're going to switch back to the config.yaml file and we're going to place it right here over the Google API key then save the last thing we need is the search engine ID and you're going to get that from this programmable search engine website it's kind of weird that you need this but you do and I'm going to drop all these links in the description below so from here click get started I already have one super AGI but if you didn't you're going to click add you can add a search engine name and then just create it it's pretty straightforward but we're going to use our existing one super AGI right there and then we need this search engine ID I'm going to copy it we're going to switch back to the config.yaml file and place that right there click save now we we have everything we need now to install super AGI we need to do two things with Docker the first thing we need to do is type Docker Dash compose up Dash Dash build and then hit enter and that's going to build everything from our repository and again you do need Docker up and running for this to work the nice thing about Docker is you're likely not going to run into any of the Python versioning issues the module missing issues that I know so many of you ask about I run into that every single time so it's really nice that the engineers put this together okay this is what it looks like when it's up and running then the last thing we've got to do is just go to localhost so go to your browser type localhost colon 3000 and then hit enter and there it is here's super AGI this is what a fresh install looks like when you first get started so we're going to go over to this agents tab right here we're going to click create agent I'm going to say this is a test agent we do need a description so testing and then we can add one or many goals here so we're going to say research AI topics for YouTube videos that's the only goal that I'm going to give it and if you have gpt4 API access you can select it here but if not you can always go with GPT 3.5 turbo so I'm going to go with turbo for now because it's really fast and here is where you select the tools so these are the tools that super AGI has to use let's look at them read email send email append file delete file the one that we definitely want is Google search read file and write file everything else I encourage you to play around with and find what works for you it also gives us some Advanced options but I'm not going to go through that now and the nice thing is super AGI seems to figure out what tools it needs as it goes then we click this button create and run and it starts there it's thinking it has all of the details of this agent and again the nice thing is you can have multiple agents running in parallel now if you want to see a little bit more information about what's going on switch back to visual studio code and you can actually look through the logs of what it's doing alright there it goes so I am a test agent my role is testing it gives a whole bunch of information I think this is all the prompts and if we scroll down here's the next thing it's going to do determine which next command to use and respond using the format specified above thoughts to achieve my goal I should start by performing a Google search to find relevant AI topics here's a Google search returned and then it gives us a bunch of information returned from the Google search and it'll continue on here's the next thought I should save the relevant information to a file for future reference plan save the AI topics to a file so it automatically grabs the tool write file and it wrote it to a file

called AI underscore topics dot txt really really cool and it thinks it's finished since I have completed my goal of researching AI topics for YouTube videos I should use the Finish command it also includes criticism just like Auto GPT does so it's thinking and then also criticizing its own thoughts and planning around those criticisms so that's super AGI now you know how to install it and run it play around with it let me know what you think if you have any questions feel free to jump in our Discord I'm happy to answer questions there and let me know what you build with super AGI if you like this video please consider giving a like can subscribe and I'll see in the next one