How SourceCrit Works
Every claim passes through our 4-agent MetaCrit pipeline. No shortcuts. No hallucinations. Just verified research.
Stage 1
Reader
Ingests the claim and identifies all factual assertions, data points, and cited sources for verification.
Stage 2
Monitor
Cross-references each assertion against 12+ authoritative sources. Flags contradictions and missing context.
Stage 3
Critic
Applies formal logic analysis. Identifies logical fallacies, survivorship bias, cherry-picked data, and false equivalences.
Stage 4
Synthesizer
Produces the final research brief with verdicts, confidence scores, and actionable recommendations.
Live Example: FDE Career Brief
A viral claim stated “FDE roles pay $180K+ and you can land one in 3 months.” Here's what our pipeline found:
Reader: Extracted 4 factual claims about FDE salaries, timelines, and market demand.
Monitor: Cross-referenced Levels.fyi, Glassdoor, LinkedIn job postings. Found median FDE salary is $145K, not $180K+.
Critic: Identified survivorship bias in “3-month window” claim and a “90-day playbook” fallacy. Only top 5% achieve this timeline.
Synthesizer: Verdict: Partially Supported. Salary claim inflated by ~24%. Timeline claim rated Fallacious due to survivorship bias.
How We Compare
| Feature | SourceCrit | ChatGPT | Perplexity | Manual |
|---|---|---|---|---|
| Source tracing | ||||
| Fallacy detection | ||||
| Multi-source cross-ref | ||||
| Structured verdicts | ||||
| Turnaround < 5 min | ||||
| Confidence scoring | ||||
| Contradiction surfacing |
Ready to verify a claim?
Submit any AI-generated claim and get a research brief in minutes.
Start Research