AI bro detector

For when "AI will replace you" deserves a second opinion

One post at a time for the most accurate reading. Text content works better than images.

AI bro energy

—

What you're doing here

Paste a suspicious post, thread, caption, or screenshot — the kind of stuff that sounds like “I built a $10M AI agency before breakfast.” Hit Analyze. You get an AI bro energy score (0–100), plain-English reasons, and the exact lines that set off the alarm. Lower is more grounded. Higher is more LinkedIn thought-leader energy. Very grounded posts score 0–10 and may show green “what checks out” chips.

How the score is calculated

An AI reads what you pasted (screenshots are analysed via OCR text only, not the image). It's trained to spot rhetoric, not to fact-check your uncle's crypto thread.

High meter — hype, fearmongering, course/DM funnels (“Comment X I'll send…”), fake “research,” claims that ignore adoption/cost for the domain
Low meter (0–10) — tool framing, steps, named studies/news, market-aware limits — not engagement farming
Summary — short neutral recap of what the post argues before the score

Very short OCR from a screenshot gets a minimum score (~25).

This is rhetoric + model judgment, not a court of law. Market notes are indicative, not verified data.

The bro tokens thing

This runs on real API calls (they cost money). So each browser gets 5 bro tokens per rolling 12 hours — not unlimited, not a ChatGPT subscription, just five full scans. The counter at the top of the input field shows how many tokens you have left.

Hit zero? Wait until your oldest scan falls outside the 12-hour window, or go touch grass. VPN won't mint extra tokens — the limit is per IP, best-effort. Shared Wi‑Fi means shared patience.

Bonus: the reaction GIF

The meter comes with a GIF that gets more judgmental as the score climbs — Stanley at the very low end, Doakes, then Batman slap at the peak. You're welcome.