They rewrite reality.
Name it. Answer it.
You don’t win the fight. You stop doubting yourself.
Paste the message from your partner or ex. Unbluff detects gaslighting, guilt trips, DARVO, and silent treatment — then writes the calm, grounded reply you can send.
No shouting. No insults. Just the reply that ends it.
ChatGPT will suggest you “communicate openly.” It will not tell you that the message is DARVO or a guilt trip — because to ask, you already have to suspect it. Unbluff names the pattern first, then gives you the grounded reply.
Three real messages, decoded
“After everything I've done, this is where you draw the line?”
Attaching a moral cost to a normal limit, so keeping it feels like an act of cruelty.
I hear that this is hard. My answer on [thing] is still no — and it isn't about [what they've done].
“I can't believe you'd accuse me of that. Do you know how that makes me feel?”
A three-step redirection. By the end of the exchange you're apologising for the complaint you made.
I'm not attacking you. The thing I raised is [fact]. I'd still like an answer on that.
“You're being dramatic, it's really not a big deal.”
Disqualifying the complaint rather than answering it. If your reaction is the problem, their behaviour never is.
It matters to me. I'd like to talk about [specific thing] — can we?
What it costs
3 decodes for $4.99 · your first decode is free · one-time payment · no subscription · decodes never expire · money-back if the decode is useless
Work messages are sensitive, so this isn’t small print. Your text is sent over an encrypted connection, analysed once, and kept only so you can reopen your own report. It is never used to train models, never shown to anyone else, and never tied to your employer. You can run your first decode without an account at all.
One engine. Different battlefields.