FeedHacker

FeedHacker cleans up your LinkedIn home feed. For each post it decides whether the post matches one of your enabled filters and, if so, collapses it to a small stub (with a "Show anyway" button) or hides it entirely. Filters you can mute: • AI slop — posts that read as AI-generated, scored by a structural algorithm that learns from your corrections (not just a word list) • Promoted posts • Company / brand posts • Newsletter signups • Hiring posts • Reaction reshares • New-job announcements • Work anniversaries • Training & certification AI slop has its own on/off plus a sensitivity slider, because it is a learned model rather than a fixed rule. Every other filter is a single toggle: on hides that kind, off doesn't. Add your own word/regex/hashtag/company filters, mute specific authors, and see exactly what has been hidden and which filter caught it. Everything runs locally in your browser. FeedHacker does not collect, transmit, or sell any data. It is unofficial and not affiliated with LinkedIn. ——— WHAT'S NEW v0.9.1 Insights now shows **which filter** hid your posts, not just how many — a per-kind table for the last 30 days, noisiest first. Buttons on a hidden post no longer need a second click: the background re-check now gets out of the way the moment you press a control, instead of a moment after. v0.9.0 Fixes a bug that was quietly making AI-slop detection worse. FeedHacker was re-checking the whole feed every 1.5 seconds and counting the same handful of posts over and over, which taught it that the clearest sign of AI writing was ordinary so it started catching less. A post is now judged once, and this update resets the AI's learning for you automatically, so detection starts again from the shipped algorithm. Your mute settings, muted authors and custom filters are untouched. Hidden posts are also no longer lumped into one row when they were hidden for different reasons, and you can now tell FeedHacker it missed one: posts it decides to show carry a small slop button that hides the post and teaches the model in the same click. v0.8.0 Every filter is now a simple on/off, and the popup is regrouped to match AI slop at the top with its own toggle and sensitivity slider, since it is a learned model rather than a fixed rule, and the plain post kinds below under Also hide.