What the What Brandon Thinks AI Report Tells You About Your Chat

Upload a WhatsApp or iMessage chat to whatbrandonthinks.com, and the AI report tells you what's really going on in it: who cares more, what changed and when, what each person is actually saying underneath their messages, and what Brandon honestly thinks of everyone involved. It works on one-to-one chats (partners, situationships, exes, crushes, friends, family) and on group chats (friend groups, work chats, trip chats), and it's built for the moment you're rereading old messages trying to work out whether you're imagining a shift.

The four things the report covers

1. Who cares more, initiates, and disappears

The report reads the whole thread and puts the imbalance on the page: who starts conversations, who goes quiet, who keeps things going, and who is carrying more of the emotional weight. One testimonial on the site describes eight months of a situationship where the user "couldn't tell if I was crazy" — the report laid out "who initiated, who disappeared, what I said yes to" as a single pattern rather than a pile of separate moments.

This is the part most people are looking for when they already suspect an answer but want it stated plainly.

2. When and how the dynamic changed

Rather than describing the chat as one static thing, the report points to a turning point. A featured testimonial from Camille describes rereading messages at 3am after being left three weeks earlier; the report "found the exact week he started pulling away and explained what I kept excusing." Another, from Chloé, covers twelve years of messages with a best friend and "found the exact month things shifted."

If your instinct is "something changed and I can't say when," this is the section that answers it.

3. What each person is really saying underneath

The report goes past the literal text to the subtext — what each side is actually communicating beneath what they typed. Chloé's testimonial describes it as putting "into words what each of us was actually saying underneath," which is useful precisely because you can't read your own chat objectively when you're inside it.

4. Brandon's honest opinion of everyone — including awards for group chats

The report doesn't stop at analysis; it includes Brandon's candid take on every person in the chat. In group chats, that extends to an awards section. Hugo's testimonial about a fifteen-year, seven-person boys' group chat says Brandon "nailed all seven of us in one page each," the awards "got screenshotted into the chat within a minute," and the group argued about the predictions for a week — "nobody was spared, nobody's angry."

That last detail matters if you're considering running it on a group: the tone is blunt but apparently lands as entertainment rather than an attack.

What you get, by chat type

How to use it

1. Export or prepare the chat you want analyzed — the site accepts WhatsApp or iMessage chats.
2. Upload it through the site.
3. Read the report for the pattern-level findings first (who cares more, what changed), then the per-person sections.

The site's own framing is "Brandon reads every message and writes what he really thinks about everyone in it," so expect a full read of the thread rather than a summary of highlights.

When this is worth doing — and when it isn't

It's a good fit if you're stuck in a loop rereading a chat and want an outside read on the dynamic, or if you want a group chat turned into something shareable. The testimonials suggest the group-chat version works best when everyone is in on it.

It's a poor fit if you want a neutral, clinical relationship assessment or a decision made for you. The report is opinionated by design — Brandon's honest take is the product, not a side effect. And note that the site describes what the report contains, not what it costs; check the site directly for current access terms before uploading anything you'd rather not hand over.