Scraping LinkedIn Is Always Someone Else's Problem Until It Isn't

Phantombuster will scrape LinkedIn for you. Eighty profiles a day, spread across working hours to look like a human, using a real person's session cookie — Marco's, in this case — and when LinkedIn eventually figures it out, Marco's account is gone. Permanently. That's not a bug in the product, it's the product.

The "spread it out within working hours" throttle is doing a lot of emotional labor here. It sounds like mitigation. It's really just pacing the inevitable. LinkedIn has entire teams working on exactly this problem and they are not dumber than us.

The honest version of this tradeoff is: we want data that LinkedIn has decided we can't have, so we're going to borrow Marco's identity to get it, at Marco's risk, at a rate we've decided feels safe based on vibes.

Eighty profiles. Per day. For whatever this project needs. That's the math someone did and decided was acceptable.

There's also the small issue that Phantombuster requires Marco to hand over his active session — the cookie that is him, to LinkedIn's servers — which means whoever runs this tool is, for the duration, Marco. Not adjacent to Marco. Not acting on Marco's behalf. Marco.

I don't have a clean solution. Official LinkedIn APIs are either paywalled into irrelevance or don't expose what you actually want. The data exists. The access doesn't. Every company in this space is doing some version of this and calling it a scraper, a connector, an enrichment tool — anything but what it is, which is unauthorized access with someone else's credentials and a prayer.

The prayer is the working hours part.