Why can Perplexity’s API not return LinkedIn URLs while the interface can?

Why can Perplexity’s API not return LinkedIn URLs while the interface can?

I would like to use the Perplexity API for LinkedIn sourcing.
The idea is simple: give an instruction such as:

“Find me a LinkedIn contact, {job post} at {company}, and return only the profile URL.”

Findings

  • Via the Perplexity interface (site/app):
    It works. The tool finds profiles, provides a hyperlink to the LinkedIn profile, and sometimes even additional contact information.

  • Via the Chat Completion API:
    Doesn’t work.
    No matter how I phrase the prompt (I tried multiple variations), the answer is always a tutorial like:
    “Open your browser, go to Google, search for ‘LinkedIn’…”

  • Via the Search API:
    Same issue. Impossible to get any result coming from LinkedIn. I only have some other websites like blogs.
    I even tried dork-style queries like:

    site:https://www.linkedin.com/in/
    
    

    → No usable output.

My questions

  1. Why this difference between the Perplexity interface and its APIs?

    • The interface seems to access LinkedIn results without a problem.

    • The APIs, however, return absolutely nothing.

    • So this is probably not a LinkedIn crawler block, but rather a restriction voluntarily imposed by Perplexity?

  2. How does the Perplexity Search API actually work?

    • Is it a headless browser like automated Chrome?

    • Or an integration with a service such as SERP API?

    • Are dork-style queries unsupported?

    • And if it’s using SERP API, is it them who block access to LinkedIn?


:backhand_index_pointing_right: Has anyone already dug into how Perplexity’s Search API works?
:backhand_index_pointing_right: And more importantly, is there any official way to get LinkedIn results through the API (as you can through the interface)?

Thank you for your help.

3 Likes

Hi, I am running into the same issue. Did you find any solutions for this? Really frustrating because the web interface returns the correct information but not the Chat api

Linkedin’s robot.txt disallows most AI bots and infact even AI user interfaces. It just looks like the user interface are working around some compliance loophole which is riskier to exploit with APIs, used by businesses. Also, Gemini’s UI actually does obey this and does not crawl linkedin (https://www.linkedin.com/robots.txt)