Installing and using Auto-GPT: An Autonomous AI Tool

• Sulaiman
• . April 18, 2023
• . Update: April 19, 2023 8:37 pm
• 419 Views
• Shares

ChatGPT has revolutionized the way we view computing. With data and natural language processing, developers are now exploring a breakthrough in AGI (Artificial General Intelligence): Autonomous AI Agent. Auto-GPT is one of these tools that enable LLMs to independently think, plan and take action. By avoiding human input, Auto-GPT allows users to accomplish their goals with Artificial Intelligence.

To install and understand how this tool works, just follow our article!

1. The heavy lifting is done by OpenAI APIs on the cloud, so you don’t need a powerful computer to install and use Auto-GPT.
2. Auto-GPT is much better at thinking and coming to conclusions if you can access the GPT-4 API. It is also less prone to hallucinations as well. To join the waitlist for GPT-4 API access, click here. However, you can also use the regular OpenAI API with the GPT-3.5 model if you do not already have access.
3. When your free credit limit is exhausted, you must add billing details and set spending limits for Auto-GPT.
4. Auto-GPT can connect to the internet and access any webpage, so you can use it to capture information from any webpage.

Step 1: The first step is to install Python and Pip

1. Install the latest version of Python for your platform from here.
2. Click on “Install Now” after enabling “Add python.exe to PATH” on Windows. You can check out the in-depth tutorial for installing Python and Pip on Windows here.
3. Run the below commands in Windows Terminal to test Python’s installation. Both commands should return the version number. Linux and macOS users should use python3 and pip3 instead.
   

   Read | World Heritage Day 2023: How Can We Preserve Heritage Sites?

Step 2: Add API keys to use Auto-GPT

1. Go to this link to open the latest Auto-GPT GitHub release page. Click on “Source code (zip)” to download the ZIP file.
2. Extract the ZIP file and copy the “Auto-GPT” folder.
3. Next, paste the folder in the location of your choice
4. Using Notepad or another text editor, open the “.env.template” file in the folder.
5. Next, go to OpenAI’s website and copy the API key. If you don’t already have an account, sign up for a free account and click “Create a new secret key”.
6. You can refer to the below image to get an idea of how to paste the API in the “OPENAI_API_KEY” section.
7. Open pinecone.io (visit) and create a free account. This will allow the LLM to retrieve relevant information from memory.
8. Click on “API Keys” in the left sidebar and click on “Create API Key” in the right pane.
9. Click “Create Key” and give the key a name like “autogpt”.
10.  Click on the button to copy the “Key Value”.
11. Then paste it next to PINECONE_API_KEY in the text editor.
12. Copy the value under “Environment” as well.
13. Paste it next to PINECONE_ENV.
14. Save the “.env.template” file.
15. Then rename the .env.template to just .env.

You’re almost done.

Step 3: is to install Auto-GPT on Windows, macOS, and Linux

1. After adding the API keys, it’s time to run Auto-GPT. Right-click the “Auto-GPT” folder and choose “Copy as path”.
2. Now open the Terminal and type cd, add a space, and paste the path you copied above. You can right-click the Terminal to paste the path quickly. Press Enter to move to the Auto-GPT folder. Right-click on the folder for macOS and Linux users to open Terminal.
3. As soon as you are here, run the below command, which will install all the required libraries for Auto-GPT.
4. Run the following command to start Auto-GPT.

Step 4: How to use Auto-GPT on your PC

1. Auto-GPT will ask you to name the AI when it runs for the first time. For example, I am creating Shopper-GPT, which finds products on Amazon. If you don’t want to create an AI with a specific use case, you can leave this field blank and hit Enter. By default, Entrepreneur-GPT is loaded.
2. The next step is to define the AI’s role.
3. Set each goal for the autonomous AI one by one. Here you tell it what you want it to do. You can instruct it to save the information in a text or PDF file. You can also instruct it to shut down once it has retrieved all the information.
4. In this step, Auto-GPT will start thinking. It will ask you to authorize actions. Press “y” and hit Enter to confirm. It will connect to websites and gather information.
5. In the fifth stage, you can read what the AI is thinking, reasoning, and planning. It also provides criticism (kind of negative prompts) so that it comes up with the right information. Finally, it takes action.
6. You can run the AI continuously without requiring user authorization by pressing y -n and pressing Enter. Make sure you replace n with a number. If you enter y -5, it means it will not ask for your permission for the next five actions. You can stop continuous actions using the shortcut “Ctrl + C”.
7. Furthermore, Auto-GPT may open Chrome automatically to gather information from the internet, as shown in the image below.
8. As I used the regular OpenAI API and the GPT-3.5 model, the model kept running in an endless loop. Auto-GPT was unable to create the file for some reason. The final information can be found under the Auto-GPT folder after you complete the task. Once you are there, open “auto_gpt_workspace”

Automate tasks with AgentGPT (Easy Solution)

Here is how AgentGPT works if you don’t want to set up Auto-GPT locally and want an easy-to-use solution for automating and deploying tasks. It’s based on Auto-GPT, but you access it directly through the browser.

1. Click the “Settings” link in the bottom-left corner of agentgpt.reworkd.ai (visit).
2. Add your OpenAI API key here. If you don’t have access to GPT-4 API, select “get-3.5-turbo” as the model and click “Save.”
3. The next step is to name your AI agent and set your objective. Click “Deploy Agent” for the Auto-GPT AI to begin processing your input.
4. To get the final result, click “Save” or “Copy”.
5. Click “Stop agent” if you want to stop the AI agent

The Auto-GPT API allows you to create multiheaded tasks and accomplish them using artificial intelligence. My testing showed that Auto-GPT failed to complete many tasks and was in an endless loop — probably because I used the ChatGPT API (GPT-3.5). The GPT-4 API is known to work best with Auto-GPT. Since the URLs were made up and throwing 404 errors, it was frequently making up URLs and couldn’t retrieve content.

Even so, Auto-GPT has a lot of potentials, and with further development, users will be able to achieve extraordinary results. In the meantime, if you want to use Microsoft JARVIS (aka HuggingGPT), follow our tutorial and learn about its insane capabilities. Additionally, we have a tutorial for using ChatGPT-based CodeGPT in VS Code. If you have any other questions, please let us know in the comment section below.