Our Methodology

How Parse Separates
Fact from Fiction

Traditional fact-checking asks: "Is this true or false?" Parse asks: "What is this article trying to make you believe, and what techniques is it using to get you there?"

The Problem We Solve

Most misinformation isn't outright lies—it's framing, omission, and manipulation. A factually accurate article can still be deeply misleading. Headlines can be technically true while conveying a false impression. Sources can be real but cherry-picked. Statistics can be accurate but presented without crucial context.

Human bias makes us vulnerable to content that confirms what we already believe. We're more likely to share emotionally triggering content without verification. We struggle to notice what's missing from a narrative because we can only evaluate what's in front of us.

Parse provides the adversarial analysis your mind can't do alone—systematically examining every claim, technique, and omission to reveal how an article is trying to shape your beliefs.

The Analysis Pipeline

Nine Specialized Agents

Each article passes through nine independent analysis stages. Every agent examines the content from a different angle, ensuring comprehensive coverage of potential issues.

PIPELINE ENTRY POINT

Extraction Engine

Deconstructs the article into analyzable components—separating claims from rhetoric, identifying sources, and categorizing content type.

What It Looks For

Article metadata (title, publication, author, date)
Individual claims and assertions made
Sources cited (named, anonymous, or none)
Article type (news, opinion, analysis, editorial)
Narrative structure and flow

Sample Output

PARSE OUTPUT

Claims Identified:12

Sources Cited:3 named, 2 anonymous

Article Type:Opinion/Analysis

Primary Topic:Economic Policy

Why It Matters

You can't analyze what you can't see. Extraction separates the article's building blocks so each can be examined independently.

Steel-Manning Analysis

Constructs the strongest possible version of every perspective presented—including viewpoints the article opposes.

What It Looks For

Core arguments from each side
Legitimate concerns being raised
Valid points that may be poorly expressed
Common ground between positions
Nuances being glossed over

Sample Output

PARSE OUTPUT

Perspective A:At its strongest, this argument suggests...

Perspective B:A fair critic would argue that...

Common Ground:Both sides agree on...

Why It Matters

Before critiquing any position, we ensure we understand it at its best. This prevents straw-manning and reveals when articles attack weak versions of opposing views.

Deception Detection

Identifies manipulation tactics across three categories: emotional manipulation, logical deception, and propaganda techniques.

What It Looks For

Emotional triggers (fear, outrage, tribal identity)
Loaded language and framing devices
Cherry-picked data or misleading statistics
Appeal to authority without evidence
Propaganda patterns (bandwagon, false urgency, enemy creation)

Sample Output

PARSE OUTPUT

HIGH Severity:Emotional manipulation detected

Quote:"...threatens everything we hold dear"

Technique:Fear appeal / tribal framing

Why It Matters

These techniques bypass rational evaluation. Identifying them lets you separate the message from the manipulation.

Critical Fact-Check

Verifies factual claims against available evidence, distinguishing between verifiable facts and unfalsifiable assertions.

What It Looks For

Statistical claims and data accuracy
Quoted statements and attribution
Historical references
Scientific consensus alignment
Source credibility and track record

Sample Output

PARSE OUTPUT

Claim:"Unemployment fell 40% last quarter"

Status:Misleading

Finding:Actual figure was 4%, not 40%

Why It Matters

Factual accuracy is the foundation. Even well-framed articles fail if built on false claims.

Fallacy Detection

Identifies logical errors and invalid reasoning patterns that undermine the article's arguments.

What It Looks For

Ad Hominem (attacking the person, not the argument)
Straw Man (misrepresenting opposing views)
False Dichotomy (artificial either/or framing)
Slippery Slope (unsupported chain of consequences)
Hasty Generalization (insufficient evidence for broad claims)

Sample Output

PARSE OUTPUT

Fallacy:False Dichotomy

Quote:"You either support this policy or you want people to suffer"

Issue:Presents only two options when many positions exist

Why It Matters

A compelling argument built on flawed logic is still wrong. Fallacy detection reveals when you're being persuaded by invalid reasoning.

Context Audit

Identifies what's MISSING—the omissions, alternative explanations, and perspectives the article doesn't address.

What It Looks For

Key context omitted
Alternative explanations not considered
Stakeholders whose perspective is absent
Timeframe manipulation (selective date ranges)
Framing that guides interpretation

Sample Output

PARSE OUTPUT

Omission:Counter-evidence

Detail:Article cites 3 studies supporting thesis

Missing:7 peer-reviewed studies with contradicting findings

Why It Matters

The most effective manipulation isn't lying—it's controlling what you don't see. Context audit reveals the invisible frame around the narrative.

Persuasion Intent Analysis

Analyzes persuasion techniques and rhetorical strategies, with special attention to content that may incite civil unrest or radicalize audiences.

What It Looks For

Emotional manipulation and tribal signaling
Us-vs-them framing and enemy creation
Dehumanization language and existential threat narratives
Calls to collective mobilization or physical action
Distinction between editorial voice and quoted content

Sample Output

PARSE OUTPUT

Civil Unrest Risk:Elevated

Attribution:Editorial language (not quoted speech)

Key Indicators:Existential threat language, mobilization framing

Why It Matters

Understanding persuasion tactics reveals what an article wants you to do—not just what to think. We also distinguish between reporting inflammatory speech and endorsing it.

Synthesis Engine

Combines all findings into a final Credibility Score using a weighted formula that balances evidence, logic, and integrity.

What It Looks For

Evidence quality across all claims
Methodology rigor of sources
Logical structure validity
Manipulation technique presence
Overall credibility indicators

Sample Output

PARSE OUTPUT

Evidence Quality:32/40

Methodology Rigor:18/25

Logical Structure:15/20

Manipulation Absence:8/15

Why It Matters

A single number is easy to game. Our multi-factor scoring means an article can't hide weak evidence behind strong rhetoric.

AI Assessment

Synthesizes all findings into a human-readable assessment from the perspective of an impartial, all-knowing analyst.

What It Looks For

Overall credibility verdict
Author/publication intent
Blind spots and omissions
Uncomfortable truths for all sides
Actionable reader guidance

Sample Output

PARSE OUTPUT

Verdict:Direct statement on credibility

Intent:What the publication aims to achieve

Blind Spots:What the article avoids

Guidance:What readers should do next

Why It Matters

Numbers and flags aren't enough. The AI Assessment translates findings into actionable intelligence—telling you not just WHAT is wrong, but WHAT TO DO about it.

The Formula

Credibility Score Calculation

CREDIBILITY SCORE = E + M + L + A

Evidence Quality

0-40 points — How well are claims supported by credible evidence?

Methodology Rigor

0-25 points — Are sources credible and diverse?

Logical Structure

0-20 points — Is the reasoning valid and coherent?

Manipulation Absence

0-15 points — Is the article free from deception tactics?

70-100

HIGH

Trustworthy with minor caveats

40-69

MODERATE

Significant issues, read critically

20-39

LOW

Major problems, seek alternatives

0-19

VERY LOW

Not reliable for forming opinions

The Final Analysis

What AI Thinks

Beyond scores and flags, Parse delivers a comprehensive assessment structured into six distinct perspectives.

Verdict

Direct statement about this specific article's credibility and the reasoning behind it.

Intent

What the author and publication are trying to accomplish with this particular piece.

Blind Spots

What this article specifically omits, avoids discussing, or fails to consider.

Uncomfortable Insight

The nuance that challenges both supporters and critics of the article's position.

Valid Point

The legitimate concern or valid point buried in the piece, even if poorly argued.

What You Should Do

Specific, actionable guidance for readers of this article.

The goal isn't to tell you what to think—it's to show you how the article is trying to make you think.

Transparency

What Parse Can & Cannot Do

What We Can Do

Identify manipulation tactics and logical fallacies
Detect omissions and framing bias
Provide a structured framework for critical evaluation
Surface questions you should be asking
Reveal the techniques being used to persuade you

What We Cannot Do

Determine absolute truth (we provide analysis, not verdicts)
Replace human judgment (use findings as input, not gospel)
Detect novel manipulation techniques we haven't seen
Account for information we don't have access to
Guarantee accuracy for highly specialized topics

Our Commitment

Parse analyzes the article as presented. We don't know what the author knows, what pressures shaped the piece, or what context exists outside the text. Use our analysis as one input among many. The best defense against manipulation is not a tool—it's a habit of critical thinking that Parse aims to cultivate.

Ready to See It in Action?

Submit any article and watch our nine-agent pipeline dissect it in real-time. Your first analysis is free.

Analyze an Article

How Parse SeparatesFact from Fiction

Nine Specialized Agents

Extraction Engine

What It Looks For

Sample Output

Why It Matters

Steel-Manning Analysis

What It Looks For

Sample Output

Why It Matters

Deception Detection

What It Looks For

Sample Output

Why It Matters

Critical Fact-Check

What It Looks For

Sample Output

Why It Matters

Fallacy Detection

What It Looks For

Sample Output

Why It Matters

Context Audit

What It Looks For

Sample Output

Why It Matters

Persuasion Intent Analysis

What It Looks For

Sample Output

Why It Matters

Synthesis Engine

What It Looks For

Sample Output

Why It Matters

AI Assessment

What It Looks For

Sample Output

Why It Matters

Credibility Score Calculation

Evidence Quality

Methodology Rigor

Logical Structure

Manipulation Absence

What AI Thinks

Verdict

Intent

Blind Spots

Uncomfortable Insight

Valid Point

What You Should Do

What Parse Can & Cannot Do

What We Can Do

What We Cannot Do

Our Commitment

Ready to See It in Action?

How Parse SeparatesFact from Fiction

Nine Specialized Agents

Extraction Engine

What It Looks For

Sample Output

Why It Matters

Steel-Manning Analysis

What It Looks For

Sample Output

Why It Matters

Deception Detection

What It Looks For

Sample Output

Why It Matters

Critical Fact-Check

What It Looks For

Sample Output

Why It Matters

Fallacy Detection

What It Looks For

Sample Output

Why It Matters

Context Audit

What It Looks For

Sample Output

How Parse Separates
Fact from Fiction

How Parse Separates
Fact from Fiction