can I extract page number and other

Probably use the JSON mode, it comes with a lot of extra details
https://github.com/run-llama/llama_parse/blob/main/examples/demo_json_parsing.ipynb

yup. it worked. thank you @Logan M

@Logan M i've noticed there is a parsing instruction in the playground. Is there one to add in Python? does it really help?

There is

https://github.com/run-llama/llama_parse/blob/main/examples/demo_parsing_instructions.ipynb

Some people have found it pretty helpful, I would say its a tad experimental right now

so, basically, adding one more step for the model, post parsing, to make sense of the extracted text? extra cost?

Actually it's applied during extraction, not post processing -- no extra cost