Scraping or API ?

[ulf-morys]

First things first: thank you, very usefuly & I am impressed !
The LinkedIn MCP-Server is very useful.
What I don't fully understand: why are there parts of the profile that the server can retrieve... and others (like the Skills section) that it can't ?
Using Claude as an MCP client to trigger a LinkedIn search, Claude tells me that "the LinkedIn API does not support the retrieval of the skill section."
Which left me kind of confused: I thought, the server would not call any API, but rather do "real" scraping from the website content.
And as a next step: if the MCP server does not do actual scraping, why does it need the cookie ? It wouldn't need cookies for an API call, would it ?

[stickerdaniel]

Hey, thanks for using this project. Claude can only see the mcp tools available, it doesn't know the details how the tool call data is retrieved. The scraper is scraping the LinkedIn page on your device. It is relying on a different LinkedIn scraper library that is pretty old. I already started writing a more modern implementation of a scraper but it is a lot of work. Once feature parity is reached, I will replace the current dependency

[ulf-morys]

Thanks for the reply - and you have all my attention for the upcoming changes !