Methodology
How a PixProof or brand profile actually gets made, so you can judge for yourself how much to trust any given one.
Sourcing standard
Every factual claim on a PixProof traces back to a real, linked source — GitHub activity, Hacker News/Reddit discussion, official recall or safety databases, real reviews, or a company's own published statements. Sources are categorized by type (Reddit, GitHub, YouTube, Hacker News, Product Hunt, Stack Overflow, or other) and shown inline, not hidden behind a generic "sources: 6" count. If a claim can't be traced to something real, it doesn't go on the page — that applies to the hype, the reality check, and the evidence for and against each individual claim.
What "AI opinion" means
The Verdict block on every PixProof is labeled AI opinion on purpose. The hype and reality sections are sourced facts; the verdict is PixPause's AI-synthesized read of what those facts add up to — not an independent lab test, not a first-party review, and not a numeric star rating. Treat it the way you'd treat an informed friend's opinion after reading the same sources you can also read yourself, linked right above it.
The four verdict states
- 🟢 Buy the Hype — the real evidence substantially backs the claim.
- 🟡 Question It — real truth mixed with real caveats; the honest answer is "it depends."
- 🔴 Bye the Hype — the real evidence contradicts or substantially undercuts the claim.
- ⚪ Insufficient Evidence — there isn't enough reliable evidence to call it responsibly. This is a real, valid verdict, not a placeholder — PixPause would rather say "we don't know yet" than manufacture a verdict just to have one.
Evidence strength
Separate from the verdict itself, every PixProof shows an evidence-strength read — Well sourced, Moderately sourced, or Thinly sourced — based on how many independently sourced citations actually back that specific verdict, counting both supporting and contradicting evidence. Two PixProofs can reach the same verdict on very different amounts of sourcing; this makes that difference visible instead of presenting every verdict with equal certainty.
Corrections
Verdicts get revisited, not set once and forgotten. When a verdict or a materially significant fact changes, it's logged with a date and a plain-language reason directly on the affected PixProof, under Correction History. Every correction across the whole site is also collected chronologically at /corrections/. A PixProof with no corrections listed simply hasn't needed one yet — this log is never backfilled to make a PixProof look more or less revised than it actually has been.
What we don't do
No paid placements or sponsored verdicts. No invented statistics — if a number appears on this site, it's real and traceable, the same standard applied on our About page. No fabricated "trust scores": a brand's Verdict Record on its profile page is a real count of PixPause's own PixProofs about it, not an independent rating of the brand as a whole.
Questions or a correction to suggest?
Visit the contact page — we read every message, including corrections to a specific PixProof or brand profile.