You need the text parser.
The WWW reader must be set as the parent instance to read the website.
Unfortunately the WWW reader is not documented further online.
I have just seen that it is a image. So the temperatura can be taken from this page:http://www.tutiempo.net/Estaciones/Cartagena/
the data I want is under Últimos datos registrados and is Temperatura, right know 16.1
Parsing web pages is a „bad solution“ because it is not sure that the result includes really the current data. You can never be sure that the data were consistent.
Small changes in page content by the provider within the $start and $end parts lead to errors.
…
It would be much better/safer to measure the temperature.