Live Crawling / Meta tags / Target Language

Link to Prompt

What the Prompt does

Create meta title + desc for a live crawled webpage

@aiprm-christophc I have a problem with this prompt. You must use [TARGETLANGUAGE] for a prompt - thats fine. But not in every case. I have switched that to german. Now I try to create a meta title and desc for my english website. But I forgot to switch from german to english - and I am getting the result in german. (But need english for the english website of course)

It would be great, if there would be a [WEBPAGELANGUAGE] variable as well which has the same rank as [TARGETLANGUAGE]. In most cases, the web page language is defined in the html code.

The results of the prompt (live crawling) are pretty good, but even Chat GPT 4.0 is a little bit too weak for it.


1 Like

Thanks @Georg_Franz that is a useful idea.

We could use language detection also.

What signal has higher priority?

  • the source code language
  • the detected language


Both good!

Is it possible to get the raw scraped page?

As far as I can see the html tags are stripped.

A new use case would be: get css/ js / pdf / whatever file and do things with it (32k version needed probably)