GPT-To-GPT Tutorial For Custom GPTs

[bloodlinealphaDev]

Hi all!

If you are interested in using Agents with your custom GPT’s I created a blog explaining the process with full source code and a simple example you can recreate to get started. I called this GPT-to-GPT communication, but it is essentially just creating LLM’s Agents using Open’s chat API and connecting it to a custom GPT using Action requests. It is quite neat and I know there are tons or use-cases for something like this, so check it out!!

The same blog is posted on LinkedIn and my blog so check out whichever you prefer! DM or email me with any questions.

Blog: Syntax Sunday: Your Typical Dev Shop | BloodLineAlpha Blog

LinkedIn: LinkedIn Article

[nadia.abchiche]

Hi,
I am interested in knowing more about this project. I a m thinking about a project for combining different LLM’s agents in a cooperative way.

[bloodlinealphaDev]

Great! That is what I was trying to showcase with this example, hopefully it helped give you some guidance. If you have any specific question DM or email me!

[Chowderr]

This is really amazing, I want to follow but when it comes to doing some things in the terminal I am completely lost. Is it possible you can release a video, or can I email you?

[bloodlinealphaDev]

More Detailed Steps

• The only content in the .env file will be: OPENAI_API_KEY=YOUR_OPENAI_API_KEY - where you need to replace " YOUR_OPENAI_API_KEY" with your own OpenAI API Key (https://help.openai.com/en/articles/4936850-where-do-i-find-my-openai-api-key)

• You will then need to continue on with step 7 and create a custom GPT using the configuration from: GPT_to_GPT/YourTypicalDevShop/YourTypicalDevShop.md at master · bloodlinealpha/GPT_to_GPT · GitHub

• Once you get to step 9 you will need to open a separate terminal window to install and run localtunnel
  

  

• Once the project is running locally on port 3000 and the localtunnel is running, it will give you a URL link which will need to be copied/updated in your custom GPT Action.

• Save/publish your custom GPT again and it will work as long as you keep your project running on port 3000 and the local tunnel running.
• Localtunnel gives you a new URL each time you start it so it needs to be updated each time when running locally… but if you host the project(API) live you will not need to do this…

Hopefully that helps!

[Chowderr]

Successfuly recreated: ChatGPT - YourFriendlyTechShop.

This was incredibly easy actually. Thank you for the help @bloodlinealphaDev .I can’t wait to build on top of this. Very amazing indeed. Projects like this and O1 really put a smaile on my face.

[bloodlinealphaDev]

@Chowderr

I can’t wait to build on top of this. Very amazing indeed. Projects like this and O1 really put a smaile on my face.

Glad you got it working!! That’s why I made this example!

[Chowderr]

I did send another email for a few last questions. Connected with you on LinkedIn as well.