Analysis

Manipulation Detector

Paste any message, pitch, post, or speech. The engine names the techniques it finds — coercion, scam patterns, gaslighting language, fallacies, propaganda framing, urgency, weasel wording — explains each one, and tells you what to do next.

Pattern engine v2 — runs in your browser AI deep scan — optional, 3 a day

Nothing you paste leaves this page unless you ask for the AI deep scan. The engine is the same rule set that powers Civic Filament: Persuasion Lab: it detects structure (if-you-don’t/I-will threats, pay-then-do bribes, ultimatums) as well as vocabulary, weighs each hit by how harmful the technique is, and reports its own confidence. Honest hedging (“perhaps”, “to the best of my knowledge”) is never counted against a writer. The AI deep scan reads context and omission: it sends your text, anonymously, to Persuasion Lab’s analysis service, which runs it through Claude and doesn’t save it. Three deep scans a day per visitor.

What a pattern engine cannot see ↓
    Try:
    0 characters

    Manipulation signal (0–100)

    0

    Techniques identified

    By category

    Annotated text

    What to do next

      AI deep scan

      How to read this

        Stylometry

        Active Development

        Civic Filament™: Persuasion Lab

        The full platform: an encyclopedia of hundreds of persuasion and manipulation techniques with harm classes, citations and defenses; AI analysis that reads context; roleplay to practice resisting pressure; and campaign tools with an ethics layer built in.

        Open Persuasion Lab →