FactMatrix – Structured Media Evaluation
FactMatrix is a next-generation analysis engine that goes beyond traditional fact-checking. It performs a multi-layered examination of any article or text, assessing factual accuracy, source quality, narrative framing, bias, legal context, and structural omissions. The result is a transparent, forensic-style report that reveals not only what the text claims, but how and why it presents its information.
Paste your article or upload a document below to receive a clear, detailed evaluation designed to help you navigate today’s complex information environment with confidence.
What FactMatrix Does
FactMatrix provides a comprehensive, structured evaluation of any article or document through six analytical layers:
Factual Verification
Identifies verifiable claims, uncertain statements, unsupported assertions, and internal inconsistencies.Source & Evidence Assessment
Evaluates the credibility, transparency, and reliability of cited sources, including anonymous sourcing and missing evidence.Framing & Narrative Analysis
Examines how language, structure, emphasis, and omission shape the reader’s perception of events.Bias Identification
Detects partisan tilt, selective storytelling, emotional loading, and asymmetrical treatment of opposing viewpoints.Legal & Institutional Accuracy
Reviews references to law, government processes, and institutions to identify errors, simplifications, or misleading representations.Integrity & Transparency Review
Highlights missing context, unbalanced perspectives, and areas where the text lacks clarity or corroboration.
Together, these layers create a clear, multi-dimensional picture of an article’s reliability, helping you understand not only what is written, but how and why it is presented.
Methodology
FactMatrix is designed to produce evaluations that are reliable, transparent, and resistant to common failure modes in automated fact-checking. It does not aim to be instantaneous. Instead, it follows a structured multi-stage review process that prioritises correctness over speed.
1. Multi-Layer Analysis
Each article is examined across several distinct layers: factual claims, sourcing, legal framing, narrative structure, bias indicators, evidentiary strength, and causal coherence. These layers are evaluated separately to ensure that no single reasoning pass dominates the assessment.
2. Internal Consistency Checks
FactMatrix performs systematic checks for internal contradictions, unsupported statements, and logical gaps. If inconsistencies appear at any stage, the system rewrites the affected sections before continuing. This reduces the risk of fabricated or speculative conclusions.
3. Evidence Anchoring
All factual claims are cross-referenced against the content of the article itself and the verifiable sources it relies on. If a claim cannot be supported by the information provided, it is flagged or qualified rather than filled with inference.
4. Explicit Uncertainty
When the available material does not allow a definitive judgement, FactMatrix states this clearly. The system avoids assumptions about intent, motive, or unverifiable causal relationships. This principle is central to its hallucination-resistant design.
5. Bias and Framing Identification
The analysis separates factual accuracy from framing choices. Language, emphasis, omissions, and selective sourcing are assessed for their influence on how readers may interpret the piece. This helps distinguish verifiable information from editorial or rhetorical shaping.
6. Legal and Procedural Context
Where relevant, FactMatrix examines whether the article’s legal references, procedural descriptions, or regulatory implications are accurately presented. It does not provide legal advice, but it highlights areas where legal terminology is used loosely or without context.
7. Self-Correction Mechanisms
Before the final output is delivered, FactMatrix re-evaluates earlier layers to ensure consistency across the entire analysis. If contradictions, unclear attributions, or imprecise statements appear, they are corrected in the final pass.
Why This Takes Time
A full multi-layer, self-correcting evaluation cannot be produced instantly without compromising accuracy. FactMatrix is intentionally designed to take the time required to minimise hallucination, avoid over-confident conclusions, and provide a structured, verifiable output. Typical analyses take between 40 and 90 seconds depending on complexity, length, and source density.
Why This Matters
Fast fact-checking tools often simplify, omit nuance, or rely on shallow heuristics. FactMatrix is built for users who require reliability over speed. Its methodology aligns with academic standards of review rather than automated shortcuts. The goal is not merely to judge whether something is true, but to understand why, how, and to what extent the information presented can be trusted.