Detect Toxic Communication Patterns in Text Messages
AI-powered Communication Pattern Intelligence (CPI) identifies gaslighting, manipulation, contempt, and 7 more toxic patterns—in messages you send AND receive.
Try It Free10 Toxic Communication Patterns We Detect
Built on 40+ years of Gottman Institute research, CPI detects the patterns that predict relationship failure—and the ones that signal real danger.
Criticism
Character attacks disguised as feedback. "You always..." and "You never..." statements that attack who someone is, not what they did.
Contempt
Mockery, sarcasm, eye-rolling, name-calling. The single strongest predictor of relationship failure according to Gottman research.
Defensiveness
Excuse-making without accountability. Counter-attacking, playing the victim, or deflecting with "yes, but..." responses.
Stonewalling
Emotional withdrawal, silent treatment, refusal to engage. Shutting down conversation as a form of control.
Gaslighting
Denying your reality, questioning your memory, trivializing your feelings. Makes victims doubt their own perception.
Manipulation
Guilt-tripping, emotional blackmail, conditional affection. Leveraging your vulnerabilities to control your behavior.
Threats & Intimidation
Abandonment threats, ultimatums designed to control, coercive language that removes your sense of safety.
Blackmail & Extortion
Leveraging compromising information, threatening exposure, financial coercion tied to information control.
Coercion
Removing agency through pressure, forced compliance, exploiting power imbalance to eliminate choice.
Passive Aggression
Backhanded compliments, indirect hostility, sarcastic compliance, proclaiming innocence while attacking.
How Toxic Communication Detection Works
CPI does not guess. It identifies specific, research-validated patterns and explains exactly what it found and why it matters.
Paste the message
Paste a message you received (inbound analysis) or a message you are about to send (outbound analysis). CPI works both directions.
CPI analyzes for patterns
Communication Pattern Intelligence scans for all 10 toxic patterns AND 7 healthy Gottman Antidote patterns. Results appear in seconds.
See what is really happening
CPI names each pattern found, explains why it is harmful or healthy, and provides the specific evidence from the text. No vague warnings — specific, educational feedback.
Get a healthier version
If you are sending a message, reFrame provides an R³ Framework™-aligned rewrite that preserves your feelings while removing toxic patterns.
Two-Way Detection: Both Sides of the Conversation
Most tools only check what you write. reFrame™ is the only tool that analyzes both directions—protecting you from toxic messages you receive AND helping you avoid sending them.
Inbound Protection
Paste what they said to you. CPI tells you if you are being gaslit, manipulated, or stonewalled—and validates your perception when your instincts are right.
“Your perception is valid. Trust yourself.”
Outbound Protection
Type what you want to send. CPI catches toxic patterns in your own message before they cause harm—and acknowledges when you are already communicating well.
“The pattern is the problem, not you.”
We Also Detect What Is Going Right
CPI does not just find problems. It recognizes 7 Gottman Antidote patterns—the healthy communication behaviors that build strong relationships.
Gentle Startup
Appreciation
Taking Responsibility
Self-Soothing Request
Repair Attempt
Vulnerability Expression
Accepting Influence
Frequently Asked Questions
What are the most common toxic communication patterns?
The most common toxic communication patterns are the Gottman Four Horsemen: criticism (character attacks), contempt (disrespect and mockery), defensiveness (excuse-making without accountability), and stonewalling (emotional withdrawal). Additional patterns include gaslighting, manipulation, threats, blackmail, coercion, and passive aggression. reFrame™ detects all 10 patterns using Communication Pattern Intelligence (CPI) technology.
How do I know if my partner is communicating toxically?
Signs of toxic communication include feeling confused after conversations (gaslighting), walking on eggshells (contempt/criticism), feeling unheard (stonewalling), or feeling guilty for having needs (manipulation). reFrame™ can analyze messages you receive to objectively identify toxic patterns — paste their message and get an instant CPI analysis that names the pattern, explains why it is harmful, and validates your experience.
Can AI really detect toxic communication in text messages?
Yes. reFrame™ uses Communication Pattern Intelligence (CPI), a patent-pending AI system built on 40+ years of Gottman Institute research. CPI detects 10 toxic patterns and 7 healthy patterns in text messages with research-validated accuracy. It analyzes both messages you receive (inbound) and messages you plan to send (outbound) for complete two-way protection.
How is reFrame™ different from grammar or tone checkers?
Grammar and tone checkers like Grammarly focus on language mechanics. reFrame™ focuses on relationship psychology — detecting patterns that predict relationship failure (criticism, contempt, gaslighting) and patterns that build healthy relationships (appreciation, repair attempts, vulnerability). It does not just fix your words; it coaches you toward communication that strengthens relationships.
What is an example of toxic communication?
A classic example is the difference between a complaint and criticism. A complaint addresses a specific behavior: "I felt hurt when you did not call." Criticism attacks character: "You never think about anyone but yourself." CPI detects this distinction automatically and explains why criticism is damaging — it triggers defensiveness and shuts down productive conversation.
See What Is Really Happening in Your Conversations
Paste any message and let CPI show you the communication patterns—toxic and healthy. Free to try, no account required.
Analyze a Message NowLast reviewed: February 2026