Install Auto-GPT Locally (Setup Guide + Use Cases)

So I recently put out a video about Auto GPT, and in that video I walked through, step by step, how to install Auto GPT on your computer locally. Now, in this AI world that we're in, everything moves so dang fast that that video that I put out is already outdated. The auto GPT GitHub page has moved to an entirely new page, and the way you actually run it is completely different than how I explained it in that video. On this video, I'm going to show you the new way to actually install Auto GPT on your computer and and run it locally. And then I'm also going to show you some examples and ways that others have been using Auto GPT in their workflows and getting good results so that you can actually know some really cool ways to use it yourself. So let's get into it Now. The first few steps are exactly the same as the first video. You still need to install GitHub and you still need to install Python. So this clip is going to walk you through that once again. So to install it on your computer, the first thing you're going to need is you're going to need Git for Windows. So come over to getforwindows.org and click on Download and install get for Windows. This is a tool that's going to make it really easy for you to pull GitHub repositories directly onto your computer and automatically set them up. Once you have get for Windows installed, you're also going to want Python installed. So head over to python.org come over to the Download section, click on Windows, and you're going to want the latest stable release. Now, as of recording this, the latest stable release is Python 3.10.0 11, which came out on April 5th. If you're going to be doing anything with Stable Diffusion, you're probably going to want to get the release right before this one and actually download Python 3.10.10. I've had a few issues trying to use the newest version 10.11 here with stable diffusion. And if you're into AI like I am, you're probably not going to want to just use Auto GPT. You're probably also using tools like Stable Diffusion, so might as well get the version that I know definitely works with Stable diffusion, which is 3.10.10. So go ahead and click on that version there, scroll down on the page and you're going to want to Click on Windows Installer 64 bit. Go ahead and let that download and then run the installer. Make sure that when you do install it, you click this box that says add Python EE to path. That's going to be important. And then go ahead and click Install now and let it install. All right, we can go ahead and close out of that. We're done there. And I'm going to go ahead and close the Python website. Now. You're going to want to find the directory on your computer where you're going to install it. In my case, I have this E drive here, which is purely for AI stuff that I do. I'm going to go ahead and create a new folder inside of my AI folder here, and I'm going to call it Auto GPT. You can install it anywhere on your hard drive that you want. This is where I'm going to put mine. And then I'm going to double click into this folder and then I'm going to come up here to my command bar where you can see my URLs. I'm going to click into this command bar so that it's typeable, and I'm going to type cmd. This is going to open up a command terminal here. All right, so once we have those steps completed, the next thing we need to do is if we head over to the GitHub page here, it's now at Significant Gravitas Auto GPT. I'll make sure the URL is below in the description. If we scroll down here, you're going to want to click into the documentation. If you click into the install instructions, I've noticed, notice that there's an error, it's sending you to the wrong page. But if you come down and click into the documentation and then click on installation, you get the correct page here. Now, the first step that it tells us is to copy and paste this little code here, get clone dash B stable, and then this URL here. So we're gonna go ahead and copy that, and then we're gonna jump back to this command terminal where we just set this all up here, and we're gonna paste that in just like that and, and hit Enter. And now it's going to clone everything from the latest stable release of GitHub for auto GPT. And now you'll see that if we jump back to our folder that we just created, there's a new Auto GPT folder in here. Let's double click into that folder. Just for simplicity, I'm actually going to close out of the command window again. I'm going to come to this bar up here, make it typable, type Command again. And this changed the directory for us to this new directory. Alternatively, you can type in CD space auto-GPT and that'll do the same thing. It will just jump you to that next directory. Now, if we open up our folder again here, you'll notice this file here called ENV Template. Let's just go ahead and make a copy of it. Paste that copy. So now we have two here, and in this new one, we're going to delete everything after env. So just go ahead and delete everything after that. So we have a file that just says envy. Go ahead and click yes to make sure it changes it. Now, if we right click on this, we can click on Open with and then Open with Notepad. And we see this text file here. And we're going to need to plug in our Open API key to get that OpenAI key. We're going to go to platform.OpenAI.com account API keys. Once you're in here, you're going to create a new secret key. I'm going to use this one temporarily and then delete it when I'm done with this video. So I'm just going to call this temp. Create my secret key. I'm going to copy this key here. We're going to open this text file again, and right Here it says OpenAI API key equals. And then it says your OpenAI API key. We're going to go ahead and replace this with the API key that we just copied. We'll paste that in there just like that. We'll come up, we'll click File Save, and then we can close out of this file here. Now, their documentation suggests that we run this with Docker. So we're going to need to go ahead and install Docker here. To do that, we'll just come over to docker.com, click on download Docker Desktop here. Go ahead and run the installation for Docker. We'll leave both of these checked and click ok. And now it's going to have you close and restart Windows, all right? After your computer's restarted, Go ahead and run Docker Desktop here. Go ahead and minimize this so it's running in the background. Then we're going to come back to the folder where we set up our Auto GPT already. In my case, it's this Auto GPT folder here. I'm going to go ahead and type CMD again to open up our command prompt. And if we jump back over to our Docker instructions here, we can go ahead and copy the Commands that it gives us, select docker, compose, build autogpt. Copy that. We'll paste this into our command prompt here and we'll let it run. All right, so now we have auto GPT installed and to actually run auto GPT we just simply grab this next code here. Docker compose, run RM AutoGPT. We'll come over here, we'll paste this into our command prompt, hit enter, and we are inside of auto GPT now. Now, when you run auto GPT by using this code that we plugged in here, Docker compose, pose, run RM AutoGPT. There are also some other arguments that you can add to the end here. So if you add dash dash speak to the end of it, it will actually speak all of the responses back to you. If you put dash dash continuous, it'll put it into continuous mode where it doesn't ask you if you want to continue every single time. If you put dash dash GPT3 only, it will only use GPT3. It won't spend more money using GPT4. Or you can set it on dash dash GPT4 only. And then there's also a debug mode. So if I want it to run continuously for me, for example, I could close out of my command line here, come back to the folder where we installed this, type cmd. To get back to this command window, grab this, run code here, copy this again, paste this, and then at the end put dash dash continuous. Now when it runs, it will run in continuous mode for me. You can see continuous mode is enabled. So let's just call this AI name AI Research Bot. And then we kind of describe what this bot is used for. In this case, a bot that searches the Internet for the hottest topics in the field of AI. And then we can start entering our goals that we want it to try to complete for us. Research trending news topics in the field of AI research popular AI trends create a text document with 10 AI related topics to discuss. Goal 4 shut down when the task is complete, go ahead and hit enter when it asks for goal five. If I just hit enter again, it will just start to run. You can see its thought process here. Thoughts? I will start by researching trending news topics in the field of AI reasoning. This will give me an idea of what is currently happening in the field of AI and what people are talking about. Plan Use the Google command to search for trending news topics and AI. Browse through the search results to find relevant articles. Create A list of 10 AI related topics to discuss and then it Starts working through the process. And since I have it on continuous mode, it's just going to keep on adding new tasks until it finally gets to its completion. You can see it googled and found a bunch of news articles. And then it says, I will browse through the search results to find relevant articles and create A list of 10 AI related topics to discuss. Browsing through the search results will give me a better idea of what is currently happening in the field of AI, what people are talking about. And then down here it actually provides criticism. I need to make sure that the articles I choose are from reputable sources and that the topics I choose are relevant and interesting. So it's trying to discern what's relevant and what's not. And it's also trying to discern what a reputable source is. Here it says, I will use the write to file command to create a text document and write down the five big trends in AI for 2023 that I found on Forbes. It's searching for more results now. Okay, so you can see here it wrote to a file called trending AI Topics command. Right to file return file written to successfully. My goal has been completed, so I will shut down. I've completed my goal of listing the top 10 trending AI topics and saving them to a text file. So there's no need for me to continue. All right, now since we've been running this in Docker, if I open up Docker here and I come to where Auto GPT is running and I click on these little three dots and click on View files, I should be able to come over to Home App User and I should be able to open this Auto GPT workspace and we've got this trending topics Txt file. I can save this file. Let me just throw it my documents folder real quick. I open it, you can see here's the top 10 trending topics that it found for me. The five biggest artificial trends in 2023 and 2022. I mean, it's not super helpful, but sometimes you need to be a little bit more specific with what you're looking for. And I guess I just wasn't specific enough. Now if you don't want to go through all this crazy installation process, there are two web based Auto GPT programs that work really well. One of them is called Agent GPT, which you can find at AgentGPT agency reworked AI. Give it a name, give it a goal, deploy the agent. It will run just like what we saw. The other one is called God Mode. You can Find it at Godmode Space. Same idea. Give it a goal like find the top 10 trending topics in AI. Click launch and it will run an AutoGPT to try to find the top 10 trending topics in AI. Now, I have made other videos about Auto GPT already, but I did want to point out some other interesting ways that some people have been using AutoGPT. So Sully Omar here shows off an example where he used autogpt with his goal of doing market research for waterproof shoes. Get the top five competitors and list their pros and cons. Include the price for each. Save the analysis somewhere. Once done, terminate. He has a little video here of it actually running through the process. You can see it doing all the work like we were just watching here. And eventually it outputs this result. Competitor one, Columbia, it gave the pros and cons. Competitor two, Salomon, gave it the pros and cons, Morel the pros and cons, Keen pros and cons, and the North Face pros and cons. And then it came to a conclusion. So that's one way that it was successfully used by Soli here. Ethan Mollick here used it to analyze the market for simulations. So he gave it the goal, research online learning simulations for business education. Provide pros and cons of six simulations along with prices. Provide recommendations for entering the market, produce a report and then finish it. Went through all of its processes at one point, got hung up on trying to create a whole bunch of different types of CSV files. But eventually it came out with this market analysis report with a SWOT analysis, Strengths, Weaknesses, Opportunities and Threats of Capitalism lab and gave a market entry strategy. So this was another fairly successful use of an auto GPT here. Here's another example from Zahid Khawaja here. And I'm just going to go ahead and play his video because it's real short where he describes what he's working on.

In this example, we're creating an autonomous agent that's going to conduct product research. We're asking it to look for the best headphones on the it's going to look at pros, cons and the prices and it's going to create a written report with all of its findings. This happens completely automatically. It's doing a web search to find the different headphones, different reviews from different websites, and it's going to scrape all that information, clean it up, condense it for us into a nice TXT file. As you can see here, it gives us the top five headphones with the pros, cons, and price, and it also gives us more information on the pros and cons.

Here's one from JB here. Read about recent events and prepare our podcast outline. He gives an example, it runs through all the processes. He's using a different user interface here, and by the end it actually gives him an entire outline for the podcast with the cold intro, the various topics, and everything to discuss on the podcast. Greg Eisenberg here actually has a nice Twitter thread with a bunch of different ways to use AutoGPT. For example, a customer service rep, Auto GPT can understand customer inquiries, provide support, and even suggest upsells. Social Media Manager Auto GPT can be used to manage social media accounts for business based on goals for retweets, likes, and even sales. A Financial Advisor Investing your money can be a daunting task, but Auto GPT can make it a breeze as a financial advisor. It can analyze financial data and provide recommendations on how to stay ahead of the curve. And then finally, this is a cool example from I'm probably going to mispronounce the name, but Varun Maya they said that AutoGPT was trying to create an app for me. Recognized I don't have Node, Googled how to install node, found a Stack Overflow article with a link, downloaded it, extracted it, and then spawned the server for me. My contribution I watched so auto GPTs are really powerful. Obviously they still leave a little bit to be desired, as shown by my own testing where it didn't actually give me the trends. It kind of gave me a list of nonsense that wasn't really helpful. So it seems to be kind of hit or miss. I think you need to be a little bit more detailed with your actual outcomes that you're searching for. But as the saying goes, this is as bad as it's going to get. These things are getting more and more powerful. These are open source, so people are iterating off of it, people are building on top of it, and it's getting crazier and crazier by the day. So pretty exciting stuff. I apologize if you watched my last video after it was outdated and struggled to install Auto GPT. Hopefully this newer update here gets you back on the right track. And if you enjoy nerding out about all this cool AI and future tech stuff, head over to FutureTools IO. This is my website where I curate all the cool tools. I come across all the AI news every single day that's coming out, and if it's a lot, it's a little too much for you. Join the free newsletter where I'll just send you the TLDR of the Week every single Friday with just my five favorite tools, some news articles, some YouTube videos. One way to make money with AI goes out every Friday and it's a nice quick read to give you the overview of everything that happened in AI for the week. So you can find that all over at FutureTools IE. Click on the free Newsletter button and I'll hook you up. Thank you so much for tuning in. If you enjoyed this video and you want to see more cool AI tutorials and stay up to date with all the AI news, make sure you subscribe to this channel and like this video. I'll make sure you see more videos just like it in your newsfeed. Really, really appreciate you. I'll see you guys in the next video. Bye. By.