TweetDetective

TweetDetective's core idea is proximity: instead of copying a suspect tweet into a separate detector, you install a Chrome extension, paste in an API key from the vendor dashboard, and from then on every tweet in your feed carries a small percentage in its lower-right corner showing the probability it was AI-written. For someone who lives in X, that removes a context switch that matters — a copy-paste trip to GPTZero is a different mental mode than a number you notice while scrolling. The vendor's published plan is Pro Detective at a discounted $9.99/mo, listed at $19.99/mo, with unlimited scans, detection of popular AI models including ChatGPT and Bard, constant updates for new models, and 24/7 support by email. Purchases carry a 7-day refund window, and the checkout page runs a documented 50%-off-for-the-first-50-customers promotion. The honest caveats are the ones the vendor names itself. The 97% accuracy and under-3% false-positive figures are based on longer texts and the FAQ explicitly says results 'might differ from shorter ones' — and tweets are typically short, so treat any single percentage as a hint, not a verdict. It analyzes English only and ships only as a Chrome extension, with other browsers and languages described as future plans. Nothing in the public documentation explains how the score is produced or which underlying model version powers it, so you cannot audit the number. And it is Twitter/X only: no other social platforms, no news sites, no bulk file scanning. Where it fits: a journalist verifying a viral post inline, a researcher sampling how much AI-generated text appears in public discourse, a social media manager scanning brand mentions for AI-written replies. Where it does not: developers needing bulk or API-driven analysis, non-English users, anyone outside Chrome, and buyers who expect the vendor's longer-text accuracy to hold on a 40-character tweet. One job, done inside the feed — that is the trade.