WHAT THEY  MEANT

Corporate messages, translated.

Paste what they said.
Find out what they meant

0/300
Try one

No account. No sign-in. Remove names before you paste.

Paste a message on the leftabove. We'll tell you what it actually means.

“Just circling back on this…”

Most filed

Corporate language, anonymously filed. A phrase appears here once 5 different people have sent it, and yours stays private until then.

    How this works

    You paste a message. It gets reduced to the words that carry meaning and matched against phrases already decoded - so "let's circle back" and "can we circle back on this?" count as one phrase, not two.

    Anything new goes to a language model told to say what the sender actually meant. Corporate bullshit is how far the words sat from the meaning; passive aggression is how much hostility was wrapped in politeness. They measure different things, which is why a restructuring announcement scores high on one and low on the other.

    Every phrase is decoded once and stored, so the same message always returns the same answer.

    The board

    A phrase reaches Most filed once 5 different people have sent it. One person sending the same thing five times counts once, because submitters are counted by a salted hash of the address, not by the number of submissions.

    That threshold is the point: a phrase five separate strangers have independently sent is generic corporate language, not any one person’s message. Yours stays private until it clears that bar.

    The board is all-time, not weekly, so nothing resets.

    It opens with a set we filed ourselves, so the section is not empty on day one. Real submissions rank alongside those and overtake them as counts grow.

    Translations are jokes, not career advice.

    Privacy

    No accounts, no sign-in, no cookies, no analytics, no third-party tracking.

    Messages you submit are stored, because that is what the cache and the board are built from. Your IP address is not stored - only a salted hash of it, which counts distinct people without identifying them.

    Emails, handles and names in greetings or sign-offs are rejected before anything is sent onward, and the model is asked to refuse anything else that identifies a real person. Neither filter is perfect, so don't paste anything confidential.