Hey folks! I'm developing a tool to help users optimize their LinkedIn profiles—including sections like headlines, summaries, experiences, and skills—by using AI to provide suggestions and feedback. Before I dive deeper into the project, I want to ensure I'm complying with LinkedIn's terms of service and protecting user privacy. Here's what I'm aiming for: users will input their LinkedIn profile URL, and then the tool will analyze the profile and give feedback, scores, and suggestions for improvements. I'm trying to steer clear of backend scraping, storing LinkedIn cookies, or anything that might violate LinkedIn's policies and lead to account bans. I have a few questions: 1. What's the best long-term, compliant architecture for creating such a tool? 2. Is it generally acceptable to perform user-consented, client-side data extraction (like having the user's browser access LinkedIn)? 3. How do reputable companies handle different user scenarios like desktop vs. mobile and automation vs. manual input? 4. If you've built something similar, what approach has proven stable and legal for you? Thanks for any insights on LinkedIn integrations, browser limitations, or compliance issues!
6 Answers
A browser extension might be your best bet! You could have users log in via OAuth, letting the extension read their profile data directly while they're logged in. It's wise to steer clear of backend scraping since LinkedIn is quick to detect unauthorized access. For analysis, allow the extension to extract relevant sections and send just that text to your backend for AI processing. Make sure users know what data is being analyzed, too. Automating updates without explicit user action can lead to policy violations. It's great that you’re prioritizing compliance—so many tools overlook this and end up causing problems for their users.
What unique features will your tool have that users can’t get from ChatGPT? I could just input my profile text there to get suggestions. The challenge is ensuring a clean input, but ChatGPT can identify and adjust for common issues.
You're on the right track with your concerns. LinkedIn's terms pretty much say no to any third-party tools interacting with profiles, and they're known for strict enforcement. Many existing tools either pivoted after receiving cease-and-desist letters or operate in a risky gray area. Serious companies either collaborate with LinkedIn officially or create products that don’t rely on their data. To avoid legal issues, consider flipping the model: have users input their profile text for your analysis. This way, you don’t need LinkedIn's API and eliminate the risk of policy violations—and it could enhance the user experience since they're not worried about account bans.
Did you know LinkedIn has a bunch of AI features already built into their platform? Still, if you’re looking for an API, check this out: https://learn.microsoft.com/en-us/linkedin/shared/integrations/people/profile-vanity-name-api.
Stick to using user-provided content only and frame it as an AI writing assistant rather than a LinkedIn data extractor. Anything automated, even on the client side, can fall into a gray area and might unexpectedly break.
What's the main goal of your tool? Have you tested the idea with real users yet?
Agreed, I'll definitely take that route into consideration during our discussions.