Home › Message Decoder

What did their text actually mean?

You've read it eleven times. You've screenshotted it to two group chats and gotten three verdicts. It's 11pm and you're doing forensic linguistics on the word "hey." There's a better way to read a message from your ex.

The 11pm problem

A message from an ex is never just its words. It arrives carrying every hope and fear you've got, and those do the reading for you: the hopeful read ("they miss me"), the catastrophic read ("they're erasing me"), the friend-committee's split decision. What none of those reads give you is the thing you actually need — a clear-eyed account of what the message does, separate from what you want it to mean.

How the Decoder reads a message

Paste the text into the app — with a line of context if you like — and the AI Message Decoder gives you three things:

• What the message is doing. Its function, named plainly: testing the waters, keeping a door ajar at no cost, pure logistics, a nostalgia hit looking for company, a real conversation request.
• What it does not establish. The load-bearing part. A "thinking of you" establishes that they thought of you for one moment at zero cost — not an apology, not a plan, not a decision.
• What replying would teach. Every reply is a small lesson in how this works now — instant warmth teaches free access; silence to a breadcrumb teaches that low-cost check-ins get nothing.

One honest limit, upfront

The Decoder never claims to know what your ex intends — no one can, and anything that says otherwise is selling you the hope. Messages have patterns; patterns can be named; and naming them is usually enough to turn a 2am spiral into a decision. But the mind behind the message stays theirs, and the decision about replying stays yours. Clarity over confusion — not prophecy.

A field guide to the common ones

• The "hey stranger." Maximum door-check, minimum content. It asks one question — does this still open? — and any reply answers yes.
• The meme with no comment. Contact with deniability. If you reply and it goes weird, it was "just a meme."
• The drunk 1am text. Honest about the feeling, useless as information. Feelings at 1am with a drink in them don't survive to 10am; decisions made off them shouldn't either.
• The happy birthday. The socially licensed check-in — the one day contact looks like manners. A bare "thank you" closes it politely; a conversation reopens everything.
• The "how's your mum?" Warmth routed through a safe third subject. Sweet, and still a door-check.
• The accidental like on an old photo. Archaeology isn't accidental. It says "I was deep in your profile" — which tells you about their evening, not their intentions.
• The "can we talk?" The only one that might be the real thing — and still worth a slow answer. The full triage covers it.

Private by design

What you paste is processed in the moment to generate the read — not stored on our servers, not logged, not used to train any model. No account required. The specifics live in the Privacy Policy.

Questions people ask