We require current data for crafting questions and evaluating answers accurately. Additionally, we need the capability to access the internet to obtain this up-to-date information

API
, , , ,
1

I work for an online education platform, and we often need to generate questions and answers based on the latest information available on the internet or from various online sources.

For example, let’s say we want to create a tutorial (create questions and judge the answers) on the latest version of Next.js (let’s say version 13). However, OpenAI’s data is limited to information available up until 2021, and fine-tuning the model isn’t always the most efficient way to generate up-to-date questions and answers, especially when dealing with code and technical content.

We’ve explored using OpenAI for evaluating learner submissions, and it works well for text-based answers. However, when it comes to assessing GitHub projects, we require access to the actual GitHub repositories.

In summary, we’re looking for a way to access the internet or utilize some other mechanism to generate the latest questions, answers, or GitHub repository information for our educational content. We’ve previously sought advice on this matter but haven’t received satisfactory solutions.

You may suggest a plugin too to do this.

1 Like
2

webrequest is pretty good.

If you need an api solution you would need to write a crawler.

3

Hi,

This looks like a sizable project, you say that you have r`cived advice in the past that did not work, can you specify what that was? On the face of it, it seems that you need some kind of web access method and then potentially an embedding method to store data larger than the context size. Accessing github projects is no different to text projects at a data level, it is only the access method that differs.

4

check this here I tried to access internet but no reply was satisfactory.