AI Conversation Analyzer:
Paste Their Words, See the Manipulation Pattern
It's late, you've reread the thread, and you can't tell if it's you. Paste the conversation or upload screenshots and see which of 14 documented manipulation tactics may be present, with the exact words behind each finding.
TRY AN EXAMPLE OR PASTE YOUR OWN:
β Results in about 30 seconds Β· π Zero-storage: Narclyze never saves your messages Β· β 14 documented tactics, exact quotes
What the Conversation Analyzer detects
The Conversation Analyzer reads a thread you paste in, whether it comes from WhatsApp, SMS, iMessage, a direct message, an email chain or a transcript, and looks for 14 documented manipulation tactics. They include gaslighting, DARVO (deny, attack, reverse victim and offender), love bombing, future faking, hoovering, triangulation and the silent treatment. For every pattern it finds, it quotes the lines that triggered it, explains what the pattern is, and describes the effect it can have on the person receiving it.
The report also gives an overall severity score from 0 to 100 with a low, medium or high rating, a short read on what the sender may have been trying to achieve, a note that puts your side of the story into words, and three suggested next steps. Think of it as a second pair of eyes on a conversation that left you confused, not as a ruling on what happened.
How it works, and what happens to your messages
You paste up to 3,000 characters, or upload up to 10 screenshots (JPEG, PNG or WebP) and let the AI read the messages from the images. Include both sides of the exchange so the replies make sense in context. Patterns are easier to see across several messages than in a single line. If you would rather not share names or phone numbers, replace them with labels such as "Me" and "Them" first; the analysis does not need them. Before you analyze, you can also choose what to focus on: a reality check, a tracker of promises and whether they were kept, the other person's own admissions and contradictions, or a check on whether the pattern repeats. Each finding comes with the exact quote behind it. The text or screenshots you submit are sent over an encrypted connection to our AI provider, Anthropic, which reads them and generates the result in real time. Narclyze does not store them on its servers and does not sell them. The details are in the privacy policy.
Three example messages and what they may be doing
These messages are written for illustration. They are not from real users. A single message rarely proves anything; what matters is whether the pattern repeats.
Example 1: Gaslighting
"I never said that. You're imagining things again. Honestly, I'm starting to worry about your memory."
It denies an event you remember and points the doubt at you instead of at the facts. One denial can be a real disagreement about what was said. When it happens repeatedly, especially about things you can check, it teaches you to distrust your own memory.
Example 2: DARVO
"I cancel one time and you attack me? After everything I do for you? I'm the one who's hurt here."
Your original concern disappears. The sender denies the problem, treats your complaint as an attack, and recasts themselves as the injured party. The conversation ends with you apologizing for raising it.
Example 3: Future faking
"Next month we move in together and I'll handle everything. Just stop pushing me about the details and trust me."
A large promise about the future is offered in place of an answer about the present, and asking for specifics is treated as a lack of trust. Plans that never get any more specific are the signal to look for.
What this tool cannot tell you
The analyzer only sees the words you paste. It cannot verify what happened, know the other person's intentions, or diagnose anyone, and an AI can misread tone, sarcasm and context. A harsh argument is not automatically manipulation. Use the result as a starting point, and consider talking it through with a qualified professional. If you want to understand a pattern in depth, our guides on gaslighting, DARVO and future faking explain each with examples and cite their sources.
Frequently asked questions
What does the AI Conversation Analyzer do?
It reads a conversation you paste in and checks it against 14 documented manipulation tactics, such as gaslighting, DARVO, love bombing and future faking. It returns the patterns it found with the lines that triggered them, a severity score, what the sender may have been aiming for, and suggested next steps.
Which kinds of conversations can I analyze?
WhatsApp, SMS, iMessage, social media direct messages, email threads or a transcribed voicemail. You can paste up to 3,000 characters of text, or upload up to 10 screenshots in JPEG, PNG or WebP. The AI reads the messages from the images, so clear, uncropped screenshots work best.
Is my conversation stored?
Narclyze does not store the text or screenshots you submit. They are processed in memory and sent over an encrypted connection to our AI provider, Anthropic, to generate the analysis, and they are not kept on our servers afterward. See the privacy policy for details.
Can it tell me whether the other person is a narcissist?
No. It describes patterns in the words you paste and cannot assess a person. Only a qualified professional can diagnose a personality disorder, and a single conversation is not enough to draw that conclusion.
How accurate is the result?
It is an AI reading, so it can be wrong or incomplete, especially with sarcasm, short messages or missing context. Treat the severity score as a prompt for reflection rather than a measurement.
If you are in danger or need to talk to someone now:
Spain: call 016 β free, confidential, answered 24 hours a day.
United States: call the National Domestic Violence Hotline at 1-800-799-7233 β confidential, available 24/7.
In an emergency, call 112 (EU) or 911 (US). Narclyze is an educational tool. It does not replace professional help and it cannot diagnose anyone.
More Narclyze tools
Related articles
Is It Gaslighting or Just a Disagreement?
Not every argument is gaslighting. Learn the key differences, see examples, and take a free 10-question self-check to see which one you may be dealing with.
14 Narcissist Manipulation Tactics, The Complete List
14 common narcissist manipulation tactics, from gaslighting and DARVO to hoovering and triangulation, each with an example and a way to defend yourself.
Future Faking: When Promises Keep You Stuck
Future faking is making big promises about the future to win trust or keep you from leaving. See the signs, real-sounding examples, and how to protect yourself.