I'm developing a basic Chrome extension (Manifest V3) designed to collect public information from TikTok while I manually scroll through its videos. I'm using MutationObserver and IntersectionObserver to capture details like usernames, descriptions, hashtags, and like counts. All actions are manual—I scroll and the extension just records info without any automation.
However, after a couple of hours of testing, TikTok bans my account. I was using a VPN (ProtonVPN) but I've tested with it before without issue, so I'm unsure how they detect my data collection. I've searched online but only find low-quality articles trying to sell scraping tools. I'm really confused about what might be triggering this ban and would appreciate any insights.
4 Answers
It's pretty tough to pinpoint the reason for the bans, especially since TikTok's detection methods can be pretty advanced. They might be using machine learning to flag accounts based on suspicious activity. Even if your actions are manual, the fact that you're injecting a script could be raising red flags. Honestly, my advice would be to avoid building anything that interacts with proprietary platforms directly, as that often leads to unpredictable issues.
MutationObservers are not commonly used in production for a reason—if I were trying to detect scraping, I’d definitely log a message whenever someone uses one. You might want to reconsider your approach. It's likely against TikTok's terms of service, so expect some bans. Maybe try being more discreet in your methods?
Yeah, using a MutationObserver is a sure way to get flagged. They're probably well aware of common scraping techniques.
Quick question: Can they actually access the console log? I had a lot of data printed there, and I'm curious if that's part of the detection process.
Sounds like they're treating your extension just like a bot. Even with manual scrolling, the extension's interactions might be triggering their protective measures. Maybe try a simpler approach, like reading the data directly from the HTML after a delay, which could prevent detection since it wouldn’t interact with the DOM immediately.
TikTok employs thorough bot-detection mechanisms at multiple levels to prevent scrapers. They could have site-wide scripts in place to spot usage of MutationObservers. One way they might do this is by altering the functionality of the observer itself on page load. Just something to think about!
Thanks for the insight! Quick question: Are they able to scan the console log? I had a lot of scraped data showing there, and I’m wondering if that could be an issue.
Mutation observers are integral to many frameworks, not just scraping. Those are common in handling dynamic content!