How to use scraped data with custom AI prompts

[scm_Tim]

Can you integrate AI chat with custom prompts that take advantage of the scraped data?

For example:

• say I want to find my customers pain points for a specific product/service based on the data it can scrape from Quora, Reddit, Google Search and People Also Ask.

• or say I want to generate a list of FAQs for a specific product/service based on the questions people are asking on those platforms.

Regular GPT answers based on an outdated knowledgebase but scraper GPT gets live data which is a lot more accurate and helpful

[scm_Tim]

SCM has been designed to do exactly that.

You can use the Article Creator, AI article generator.

1. You load your custom prompts
2. Insert scraped content macros into your prompts.
3. Run the task and scraped content will be used to generate prompts sent to the AI.

Eg You can take questions from Google people also ask and put them into a prompt and ask the AI writer to answer questions.

Available scraped content macros

%scraped_title%
%scraped_body%
%scraped_paragraph%
%scraped_sentence%
%paa_question%
%paa_answer%

[scm_Tim]

AI Macros Tutorial

[News_KThksBuy]

%scraped_body%

This will use exactly what is being scraped? Can we put into ai like this [rewrite the output %scraped_body%] so the ai will rewrite what is being pulled from the internet.

[scm_Tim]

@News_KThksBuy

%scraped_body%

The scraped content is coming from scraper settings below here

@News_KThksBuy

Can we put into ai like this [rewrite the output %scraped_body%] so the ai will rewrite what is being pulled from the internet.

Yes you can do this, but its not necessary if you want to re-write content better to select AI rewriter instead.

[News_KThksBuy]

Hi founder,

I’m actually still quite confused after trying out using it last night.

1. If I do not use scraped macros and only using keyword and language, does the software still look for scrape content and rewrite it into my content?

2. The scraped macros, how to use it correctly? For example I want to create a full article with 5 subheadings and using all scraped content from Google. Would you show me example how I can do that?

3. How to answer to paa using scraped content from Google?

4. This is my current settings, but the subtitles don’t seems to be unique, it keeps repeating.

[scm_Tim]

@News_KThksBuy

• If I do not use scraped macros and only using keyword and language, does the software still look for scrape content and rewrite it into my content?

Yes it will still scrape content

To turn it off set, set urls to 0

@News_KThksBuy

• The scraped macros, how to use it correctly? For example I want to create a full article with 5 subheadings and using all scraped content from Google. Would you show me example how I can do that?

If you just want to use scraped content, turn of AI writer and just use the scraped articles generator.

@News_KThksBuy

How to answer to paa using scraped content from Google?

Simple way

Macro way

Output

@News_KThksBuy

This is my current settings, but the subtitles don’t seems to be unique, it keeps repeating.

I tried it and you are correct. has repeats.

I suggest you try to make subheadings different by using scraped content

%s1%=[write subheading for %keyword% %related_keyword%]
%s2%=[write subheading for %keyword% %related_keyword%]
%s3%=[write subheading for %keyword% %related_keyword%]
%s4%=[write subheading for %keyword% %related_keyword%]
%s5%=[write subheading for %keyword% %related_keyword%]

Output

This is a stop gap measure as I am investigating how to get AI to write an article outline and use that outline to write an article using multiple prompts.

Cost vs Result? Need Help

I can convert your prompt to write a section easily [image] I am thinking of adding %outline% as a special macro. It runs this prompt automatically [article outline for keyword%] output tends to be like this: [image] Take each line one at a time from output as a section ie introduction importance of feeding dogs a balanced diet. I need to test this to see how well it works.

[News_KThksBuy]

Thanks founder for replying so quickly to my questions.