Using InfraNodus API to Generate Questions for Any Text

InfraNodus API can be used to generate interesting research questions for any discourse. For example, you can use it for a collection of research papers, current news (via an RSS feed), web pages, or Google search results. These questions can help better understand the content and see how it can be further developed in an interesting way.

To do that, you can either save an existing text as an InfraNodus graph or send the text with the API request directly. The response will be a set of AI-generated questions that you can use to develop the original discourse further. The questions generated can also be used as prompts in your own LLM agentic workflow.

Here are a few examples of how this can be set up:

Example 1: Question the news of the day using Make.Com automation (see below)

Example 2: Question your existing knowledge base with InfraNodus API (see below)

Example 3: Build a custom AI chatbot that generates pertinent research question

Example 1: Question the news of the day

In this example, we use a custom scenario we created using the Make.com automation tool to query the news of day on the topic of AI (provided with a custom RSS feed). The content of the news is fed to the InfraNodus API endpoint, which generates 3 research questions, which are then further refined by an LLM node that chooses the best one and sends it to you via Telegram:

This scenario is available and can be set up via https://eu2.make.com/templates/12339-question-the-news-of-the-day-and-send-ideas-via-telegram

As a result, you get very interesting questions that bridge the gap between different topics discussed in the news. For instance, compare the question generated by  the standard GPT-4o AI (top) and the InfraNodus endpoint (bottom). You will see that the standard AI question is very generic (in fact, it always tends to connect to the topic of ethics), while the InfraNodus one is very specific and pertinent to the news sources it analysed.

The InfraNodus response also provides the names of the topics that it is using to generate the question, revealing its logic much better than a standard LLM:

In fact, InfraNodus points our attention to regulation but making a very interesting analogy between Alibaba's business and the new robotics startups. 

Example 2: Question your existing knowledge base

In this example, we will show you how you can generate interesting questions for a collection of documents that you upload to InfraNodus. You can use our public API playground at RapidAPI (note, you will need an API key from InfraNodus which can be obtained at https://infranodus.com/api-access. 

Step 1: Add a text to InfraNodus

Start by adding a text to InfraNodus. For instance, it can be a collection of scientific articles or AI-generated knowledge graph for a certain topic. In our case, it's going to be a knowledge graph on high blood pressure and various ways to affect it through lifestyle and dietary changes. 

To do that, we

- go to the https://infranodus.com/apps page

- choose the "Brainstorming" tab

- go to AI generated > Knowledge Graph

- type in blood pressure

- generate the graph

It will look something like this:

You can already see the main topics generated based on the topical clusters:

- Fluid dynamics

- Hormonal regulation

- Cardiovascular fitness

- Vascular control

- Hemorrage response

- Nutritional impact

This underlying graph structure will be used when you query InfraNodus using the API endpoint to generate a research question. It will take into account the underlying graph structure to identify the structural gap and retrieve the clusters that are not yet well connected.

For instance, in our case, there is a gap between

- Vascular control and

- Nutritional impact

These clusters will be used to generate a research question:

This is where InfraNodus is different from all other AI tools: it actually shows you what happens under the hood. So you can reveal the logic and have a clear understanding of how it works.

Step 2: Create a POST request to the following endpoint:

In order to query this graph and to get a list of the questions, we need to form the following request to the InfraNodus API endpoint `graphAndAdvice`:

https://infranodus.com/api/v1/graphAndAdvice?doNotSave=true&addStats=true&optimize=gap&includeGraphSummary=true&includeGraph=false

the POST parameters to use:

{
    "requestMode": "question",
    "modelToUse": "gpt-4o-mini", 
    "name": "support_nodus_labs",
    "aiTopics": "true"
} 

This is telling the InfraNodus API to identify the two topical clusters that are not yet linked with the biggest gap between them and to then generate a set of research questions that bridge those gaps.

You can see an example of this access point and try it out live in Rapid API: 

https://rapidapi.com/infranodus-infranodus-default/api/infranodus/playground/apiendpoint_ee2add34-db6d-4358-ae1b-a2aba9ca8964

Note, you'll need to add the authorization token for InfraNodus API from https://infranodus.com/api-access first.

Step 3: Extract the aiAdvice response 

It is an array of generated questions that bridge the gap will be in the response body aiAdvice array. 

{
"aiAdvice": [
{
"text": "Question 1?",
"finish_reason": "stop"
},
{
"text": "Question 2?",
"finish_reason": "stop"
},
{
"text": "Question 3?",
"finish_reason": "stop"
}
],

For instance, in our case, we will have the following response:

The question here says:

How does the consumption of high-protein foods, such as soy products and fish, influence blood pressure regulation and oxygen delivery in organs through mechanisms involving angiotensin II and vascular health?

It is a very interesting question because it proposes a concrete way that high blood pressure could be alleviated through consuption of protein-rich foods that decrease the level of angiotensin II in the body and therefore lead to vasolidation (relaxation of blood vessels which lessens constriction). Normally, angiotensin II is produced when the body is in a state of dehydration or has low sodium content. Eating protein can reduce its production and thus lower the blood pressure.

Those same questions can be extracted as prompts and integrated into an LLM workflow to be used in a prompt to generate interesting conversation starters in a chat or to be sent to another, more powerful and more expensive LLM, such as ChatGPT 4.5 to explore this topic further.

Read more about it in our article on creating a custom AI chatbot with InfraNodus and Dify.