In 2026, shadowbans on X are driven by a trust and behavior score, not just one action like auto-replies. Two accounts can do the same thing and get completely different results depending on their history.
Account age and trust matter a lot more than before. New accounts are heavily sandboxed, and replies are usually the first thing to get limited. Even light automation or repetitive engagement can trigger visibility limits if the account hasn’t built enough trust yet.
Engagement patterns are also a big factor:
repetitive replies or similar phrasing
replying mainly to large or viral accounts
lots of outbound engagement with little inbound engagement
automation that looks “too clean” or perfectly timed
That’s why some people can run full automation without issues - those accounts are aged, have organic interaction, and already sit in a high-trust bucket. Meanwhile, newer or low-trust accounts get flagged for much less.
So in short: account age + trust score come first, engagement patterns second, and tools/automation only matter in how human or abnormal the behavior looks.
