Issues with HTTPS Requests in an OpenAI Model

Hello,

I am using an OpenAI model to make HTTPS requests to a custom endpoint, but I am encountering issues when attempting to connect to the server from this environment. Below are the details:

  1. Context of Use:
  • The endpoint requires an HTTPS connection.
  • The server responds correctly from other tools and environments, but requests fail when made from this OpenAI-based system.
  1. Errors Observed:
  • I encounter the following error during requests:
HTTPSConnectionPool: Max retries exceeded with URL
  • Even when disabling SSL verification (verify=False), the connection does not succeed.
  1. Tests Performed:
  • Verified that the server responds successfully from tools like Postman.
  • Tried enabling and disabling SSL verification in the requests.
  1. My Objective:
  • I need guidance on why this environment is unable to connect to the server using HTTPS and how to resolve this.

I would greatly appreciate any insights or recommendations to resolve this issue. Please let me know if you require additional details or logs.

Thank you for your support.

Best regards,
Rafael Developer of milcontratos.

OpenAI API models cannot access the internet, even through code interpreter (they plugged another hole…)

You will need to have a function built with your own code for the AI to call. You will likely want to describe use of the internet by task and usefulness, which is parsed and placed into the constructed network request, rather than “you can make random network requests for the user from our IP”.

Thank you for your response, but if ChatGPT doesn’t have internet access, how can I create a model that needs to use my API to extract data if the model itself cannot connect to the internet? The idea behind creating this model is to export it to the GPTs Store, but if it cannot connect, how can I make it work? I have been reviewing some GPTs from the store, for example, Scholar GPT, and the model can connect to sider.ai to retrieve data. I understand that sider.ai is an external service to ChatGPT, but that model can connect, whereas I can’t find a way to make my model connect to my API.

You have to distinguish if you are using ChatGPT GPTs, or the API.

In ChatGPT, a GPT can use actions, functions that can call an external API.

Here is setup documentation.

https://platform.openai.com/docs/actions/introduction

Thank you very much for the information, but I have already found the solution. Since I have a basic plan, it does not provide the advanced options needed to connect to external APIs; this is only possible with an Enterprise plan.