Over the past few weeks, we've had a ton of great news about ChatGPT, GPT-4, and more. Some of you may have seen something about AutoGPT, but I certainly don't blame you; you may think it's just another GPT plugin or Chrome extension. But AutoGPT is much more than that.
What is AutoGPT?
AutoGPT combines GPT-3.5 and GPT-4 via an API, allowing the creation of projects that iterate on their own cues, and review each iteration to improve and build on it. How exactly does this work?
AutoGPT requires:
- artificial intelligence name
- artificial intelligence role
- Up to 5 goals
For example:
- Name : Chef-GPT
- The role : An artificial intelligence designed to find common recipes on the web and turn them into Michelin-star quality recipes.
- Goal 1 : Find easy recipes online
- Goal 2 : Turn this simple recipe into a Michelin-star-quality version.
Once AutoGPT satisfies the description and goals, it will start doing its own thing until the project reaches a satisfactory level.
So what are the benefits of AutoGPT? First, it's important to note that GPT is able to write its own code using GPT-4. It also executes Python scripts, enabling it to recursively debug, develop, build, and continuously improve itself. Crazy? AutoGPT is a self-improving AI - exhibiting true AGI (Artificial General Intelligence) capabilities.
The feedback loop for AutoGPT looks like this:
- plan
- criticize
- Behavior
- read feedback
- plan
AutoGPT will read and write different files, browse the web, while reviewing and checking its own hints - just to make sure the item is what the user intended. You give it a goal, and it scrapes the best information on the web, and it automatically does it for you, constantly improving itself.
AutoGPT will ask for your permission after each prompt to make sure the project is moving in the right direction.
Below is an example of an application created by AutoGPT for computer science engineer Varun Mayya . AutoGPT realized that Varun didn't have Node, so it googled how to install Node, then AutoGPT found a Stackoverflow article with a link, downloaded and unzipped it, and then spawned the server on Varun's behalf.
How to use AutoGPT?
In order to use AutoPGT, credits will be used from your OpenAI account. However, you can use up to $18 which is included in the free version.
As I mentioned above, AutoGPT asks for your permission after every prompt, which means a lot of testing on your part. This allows you to test your AI project and make it fit your needs before spending any money.
Installation and requirements
In order to use AutoGPT, you will need:
- Python 3.8 or later
- OpenAI API secret
- GPT-4 API access
- Pinecone API Key
- ElevenLabs API for text-to-speech projects
In a CMD, Bash or Powershell window, clone the repository:
<span style="background-color:#f5f5f5"><span style="color:#111111"><code><span style="color:#000000"><span style="color:#000000">git clone https</span></span><span style="color:#666600"><span style="color:#666600">:</span></span><span style="color:#880000"><span style="color:#880000">//<a data-cke-saved-href="http://github.com/Torantulino/Auto-GPT.git" href="http://github.com/Torantulino/Auto-GPT.git" class="vglnk">github.com/Torantulino/Auto-GPT.git</a></span></span></code></span></span>
Enter the project directory:
<span style="background-color:#f5f5f5"><span style="color:#111111"><code><span style="color:#000000"><span style="color:#000000">cd </span></span><span style="color:#008800"><span style="color:#008800">'Auto-GPT'</span></span></code></span></span>
Install the required dependencies:
<span style="background-color:#f5f5f5"><span style="color:#111111"><code><span style="color:#000000"><span style="color:#000000">pip install </span></span><span style="color:#666600"><span style="color:#666600">-</span></span><span style="color:#000000"><span style="color:#000000">r requirements</span></span><span style="color:#666600"><span style="color:#666600">.</span></span><span style="color:#000000"><span style="color:#000000">txt</span></span></code></span></span>
You will then need to navigate to the folder and rename it to . Once done, open. Then you want to use your ..env.template.env.envOPENAI_API_KEY
ELEVEN_LABS_API_KEYIf you are using it for speaking purposes, you will also need to fill out yours.
How to get the key:
- OpenAI API keys are available from: https://platform.openai.com/account/api-keys .
- ElevenLabs API key from: https://elevenlabs.io .
Once this is all done and successful, you need to run in a CMD, Bash or Powershell window:
<span style="background-color:#f5f5f5"><span style="color:#111111"><code><span style="color:#000000"><span style="color:#000000">python scripts</span></span><span style="color:#666600"><span style="color:#666600">/</span></span><span style="color:#000000"><span style="color:#000000">main</span></span><span style="color:#666600"><span style="color:#666600">.</span></span><span style="color:#000000"><span style="color:#000000">py</span></span></code></span></span>
You are ready to start using AutoGPT!
Please refer to GitHub if you have any questions .
AutoGPT Demo
You can download a demo video from the Auto-GPT GitHub repository