ReachOS - Grammarly for tweet reach — scores drafts against X's algo

Score your tweet drafts against 36 rules from X's open-sourced recommendation algorithm before you post. Paste a draft, get an instant score with fixes. What-If Scenarios show impact of each change: - Remove external link: +52% reach - Add an image: +38% - Post at peak time: +25% - Optimize all rules: +77% Runs entirely in the browser. No account needed. AI features (slop detection, auto-optimize) are optional and use your own API key. Free and open source (MIT).

Hey everyone 👋 I built ReachOS because I kept wondering why some tweets reached thousands and others died silently. I wanted to see the score before hitting post. So I read through X's open-sourced algorithm (twitter/the-algorithm on GitHub). Nobody really dug into the scoring logic — most coverage was surface-level. I found 36 concrete rules: link penalties, media boosts, engagement multipliers, time decay. The biggest surprise: a single external link cuts your reach by 30-50%. Replies are worth 27x more than likes. And the first 30 minutes basically decide everything. I turned all of it into a scoring tool. Paste your draft, get a score, see what to fix. The What-If feature lets you toggle changes and see predicted impact instantly. It's free, open source, runs in your browser. No data leaves your machine. Would love your feedback — especially on which rules surprised you most.