Manipulator

About

🎯 SEE THROUGH THE ALGORITHM

Manipulator scores every tweet on X (Twitter) from 0 to 10 based on how hard it's trying to extract your attention. Engagement bait, clip farms, click-through tricks, themed-account spam — they all get flagged at a glance.

The badge sits inline next to the timestamp. Hover it for the full breakdown: which signals fired, how much each contributed, plus a profile of the poster (country flag, account age, your running history with that handle).

The score is calculated by ML model that runs in your chrome browser working along with the heuristic rules.

🔒 PRIVACY

Manipulator runs entirely in your browser. The only network requests:

• X's own API 
• Hugging Face, one-time download of the local ML model and library, then cached forever

Tweet content, account handles, profile data, and your scoring history are NEVER transmitted to me or anyone else. No analytics, no telemetry, no tracking.

📦 WHAT YOU GET

• Inline 0–10 score badge on every tweet
• Hover tooltip with full signal breakdown
• Country flag of the poster's origin
• Per-handle reputation that learns as you scroll
• Color coding: green ≤4, yellow 4–7.5, red >7.5
• Runs entirely on your machine

⚠️ LIMITATIONS

• English-language signals only — non-English tweets score less accurately
• Works on x.com and twitter.com only
• First page load is ~5 seconds slower while the ML model loads from disk
• Heuristics are best-effort — some legitimate tweets will score higher than fair, some clever bait will slip through

This is a tool for your eyes, not a content filter. It doesn't hide, block, or change anything X shows you. It just labels what's already there.