The algorithm came between people I loved.

Social media optimizes for attention, not user intent. Feed Filter is the user-owned control layer for algorithmic feeds.

I don't mean that figuratively.

A few years ago, social-media-fueled conflict helped create a rift between my sister and me. Things that felt urgent on a screen became personal in real life. That was when this stopped feeling like “bad content” and started feeling like a loss of agency.

Then I noticed it was happening to me too.

I'd open Twitter to check one thing and look up forty minutes later, upset about things I hadn't been thinking about before I opened it. Not just more reactive in a general way but specifically worse at assuming the people around me were acting in good faith. I felt informed. I was being trained.

So I did what everyone does. I muted accounts, clicked “not interested,” and blocked the worst ones. For a day or two, the feed felt cleaner. Then the same pattern came back through different accounts, slightly repackaged.

That's when the willpower frame collapsed. I wasn't weak or undisciplined. The tools I'd been given solved the surface problem, not the pattern. I was doing maintenance on a system that had no interest in being fixed.

What hurts you is exactly what the platforms profit from.

Strong reactions create engagement. Engagement creates more time on app. More time on app creates more ad inventory. That incentive doesn't care whether you leave feeling informed, manipulated, angry, distracted, or closer to what you came for.

Platforms are structurally conflicted: better user control can reduce time on app. Native tools like mute, block, and “see less often” are real, but they operate at the surface. They make users do endless manual cleanup while the underlying pattern keeps returning.

Before we built the next version, we fielded a US-national survey of daily scrollers. 771 people completed it reliably. The single strongest signal: 89% wanted the tool to catch repeat offenders automatically — the thing their platforms don't provide. The seven findings, including why blocking “comes back,” are written up in the survey results.

So I built something that does.

Feed Filter starts with a handle-based feed audit. You enter your social handle, choose a platform, and get a report showing who and what is degrading your feed: repeat offenders, clickbait and ragebait patterns, misleading claims, low-value promotional content, and words or topics to mute.

Seeing the problem clearly is the first step. The audit gives you proof before asking you to install or configure anything.

The next layer is automatic cleanup: one-click mute, block, and unfollow; recurring audits; and real-time filtering through the app and browser extension.

It won't fix what happened to my family. But it gives you something the platforms were never going to hand over by default: a feed that actually runs on your rules.

I made one decision early on that I'm not going to compromise on: nobody can pay to change how their account is classified. Not an advertiser, not a partner, not an account with ten million followers. The same rules run on everyone.

We never ask for your password or touch your private messages. Everything we check is content that's already publicly visible. And whatever your results show, they're yours. You decide whether anyone else sees them.

Feed Filter is live. 626 email signups, 559 social handles submitted, 5 paid customers, and 133,000+ social posts processed across the current product and earlier prototypes.

I personally read every email. help@feedfilter.com