Editorial Standards

Fact Storm Hub publishes hundreds of facts every month. Without clear standards, that volume could easily become noise. This page documents the standards we hold ourselves to, so readers can understand exactly how our content is created, fact-checked, and published.

Core Principles

1. Accuracy Over Speed

We’d rather publish one well-researched article than five rushed ones. Articles that don’t pass our verification process are rejected, not published with disclaimers. Speed never trumps accuracy.

2. Transparency About AI

We use AI tools (Google Gemini and Anthropic Claude) to draft articles and FLUX to generate images. We don’t hide this. Every article includes a clear AI disclosure at the bottom, and our HTML includes meta tags so search engines can recognize AI-assisted content. Transparency is non-negotiable.

3. Sources Always Visible

Every article ends with a Sources section that links back to the research used during drafting. No black-box claims. If we say “scientists found X”, we link to the source where that finding was published.

4. No Clickbait Without Payoff

Headlines should match content. If a title says “5 mind-blowing facts”, the article must deliver 5 actually mind-blowing facts. Our automated quality check enforces this: the number in the title must match the number of fact sections in the body.

5. Update or Delete

When we discover an error, we fix it publicly and note the correction. When new research supersedes an old fact, we update the article and add a “last updated” date. We don’t quietly delete posts to hide mistakes.

The Editorial Pipeline

Every article passes through 7 quality gates before publication:

Gate 1: Topic Discovery

Topics come from one of three sources:

Trending sources: Google Trends, Reddit Hot, YouTube Trending (filtered for English content)
Curated bank: Pre-approved evergreen topics that match our nine content categories
AI-generated suggestions: When trending sources have no relevant topics, our AI proposes specific story-driven ideas based on niche keywords

Topics that don’t fit our content categories or violate our content safety policy are rejected before any research begins.

Gate 2: Live Web Research

For trending topics, we run live web searches via Google Search APIs (Serper.dev) to gather verified facts. Our research system follows strict verification rules:

A fact is verified if mentioned in 2+ different search results
A fact from 1 source is acceptable only if the source is authoritative (.edu, .gov, .org, peer-reviewed, official records)
Single non-authoritative sources are excluded
We prefer current-year data over older data when both exist
We never fabricate facts not present in search results

Gate 3: AI Draft Writing

We use Google Gemini 2.5 Flash (with Claude as fallback) to draft articles. The AI is instructed to:

Anchor every claim to the verified research (not the AI’s training data)
Include specific numbers, dates, names, and sources
Use storytelling structure, not bullet lists of facts
Cite sources inline where relevant
Avoid generic phrases like “many believe” or “scientists say”
Write 1500–2500 words with 5–7 numbered fact sections

Gate 4: Automated Quality Check

Every drafted article runs through automated checks:

Fact count match: If the title says “5 facts”, the body must have exactly 5 numbered H2 sections
Word count: 800–2500 words (rejected if outside this range)
SEO score: Calculated from title, meta description, keyword density, headers
Readability score: Flesch-Kincaid grade level analysis
Title structure: Must have proper H1, H2 hierarchy

Gate 5: AI Editor Review

An AI editor (powered by Claude) reads the article and scores it from 1–10 across these dimensions:

Hook quality: Does the opening grab attention?
Originality: Does it add value beyond what Wikipedia says?
Source diversity: Multiple credible sources cited?
Brand voice: Consistent with Fact Storm Hub’s editorial style?
Content safety: No divisive, inflammatory, or harmful content?

The editor assigns a tier:

Tier A (8–10): Publish immediately. Eligible for YouTube video version.
Tier B (6–7): Rejected. Sent back for rewriting up to 2 times.
Tier C (1–5): Discarded entirely.

Only Tier A articles (score 8 or above) get published. This is intentional — we’d rather publish less but better.

Gate 6: SEO & Internal Linking

Approved articles go through an SEO optimization step:

Meta description rewritten for click-through rate
Internal links added to related published articles (no broken links)
Featured image generated via Replicate FLUX (style matches niche)
YouTube SEO tags suggested
Schema.org structured data added

Gate 7: WordPress Publishing

Final step: the article is published to WordPress with:

Featured image
Category & tags assigned
FAQ schema for rich snippets
Sources section appended
AI disclosure footer added
Yoast SEO data populated

Content Safety Policy

We explicitly do not publish content that:

Promotes hate speech or discrimination based on race, religion, gender, or nationality
Spreads misinformation about health, vaccines, or medical advice
Endorses violence or political extremism
Targets specific individuals for harassment
Includes adult content, gambling, or substances
Could cause communal division or religious conflict

Our AI editor explicitly checks for these issues and assigns Tier C (rejected) to any article that fails the safety check.

Refresh Policy

Older articles are periodically refreshed with new research data. About 17% of our publishing cycles allocate effort to refreshing instead of creating new content. This ensures:

Statistics stay current (e.g., a player’s career stats update over time)
New discoveries get added to relevant older articles
Broken links are caught and fixed
SEO rankings improve as Google sees fresh content

Corrections Policy

When we make mistakes, we:

Acknowledge the error publicly within 48 hours of credible feedback
Update the article with the correct information
Add a correction note at the top of the article (if material)
Update the “last reviewed” date

We do not quietly delete or rewrite articles to hide errors. Transparency builds trust.

What We Won’t Do

Pay for backlinks. Our SEO grows organically or not at all.
Republish copied content. Every article is original drafting based on our research.
Hide AI usage. We’re transparent about using AI tools in our editorial process.
Publish content we wouldn’t put our name on. If it doesn’t pass the editor’s score, it doesn’t get published.
Take political sides. We focus on facts, not opinions.

Questions or Concerns?

If you have feedback about our editorial standards, want to report an error, or think we should change our policies, please contact us. We take editorial integrity seriously and welcome reader input.

Last updated: April 2026