It's a personal project — me plus an AI pipeline I wrote that pulls mold-related news, research, and regulation every day, runs each story through a stack of AI checks, and posts what holds up.
Mold illness affects millions of people, but good information about it is buried. Research papers sit behind paywalls written in language nobody outside academia can parse. News coverage is scattered across local outlets that never connect the dots. And the stuff that does go viral is usually wrong, overhyped, or trying to sell you something.
I wanted to fix that. So I built a pipeline that gathers mold news, research, and regulation every day, runs it through a stack of AI checks, and posts the stories that hold up. Every article is rewritten for clarity, checked against medical compliance rules I baked into the pipeline, and verified for scientific accuracy before it goes live.
No paywalls. No login gates. No ads disguised as articles. Just the news, the research, and the context you need to stay informed about mold exposure and your health.
I run growth at MoldCo, a clinician-led virtual clinic focused on mold toxicity treatment. After spending months watching good mold research get zero mainstream attention while misinformation spread unchecked, I decided to build something on the side. The Mold Report started as an experiment: what if you gave a stack of AI agents the same standards as a real newsroom and pointed them at mold science? Turns out, they're pretty good at it. I handle the product, the strategy, and the occasional "please stop writing headlines with semicolons" feedback to the bots.
Every article runs through a stack of AI checks I built. Calling them a "team" is a stretch — they're scripts with prompts. But each does one job and does it consistently. Here's who does what:
First line of defense. Runs three detection layers — URL matching, title similarity, and content fingerprinting — to catch the same story arriving from different sources. Caught and killed 14 duplicates on day one.
Reads every article and scores it 1-10 on newsworthiness. If it doesn't hit a 7, it doesn't get published. Ruthless but fair. Rejected an article about "mold-inspired art installations" without hesitation.
Screens reader-submitted tips for editorial validity and scientific alignment. Great at spotting spam, self-promotion, and that one guy who keeps submitting links to his ozone generator website.
Rewrites titles so actual humans want to read them. Turns "Combined toxicity prediction of deoxynivalenol and fumonisin B(1) by physiologically based toxicokinetic modelling" into something you'd actually click on. You're welcome.
Takes raw source material and rewrites it in plain, clear language. Trained to write like a smart journalist, not like an AI that just discovered em dashes. Keeps every summary between 150-250 words.
Makes sure nothing we publish makes a medical claim it shouldn't. Cross-references every article against MoldCo's compliance rules. Obsessed with the difference between "treatment" and "recovery." We let them be.
Checks science articles against Shoemaker Protocol research and evidence-based mold literature. If a study's methodology is questionable or the claims don't hold up, this agent flags it before publication.
Writes the meta titles and descriptions that search engines see. Keeps titles under 60 characters and descriptions under 160. Gets unreasonably invested in whether "mold exposure" or "indoor mold" has better search volume.
Finds the right image for every story. Pulls Open Graph images from source articles or assigns category-specific visuals. Won't stop asking when we're getting a real photo library.
Heads up: I'm Julien Lepleux. I built and run The Mold Report on the side. My day job is leading growth at MoldCo, a clinician-led virtual clinic for mold illness — so yes, I have a horse in this race.
What this means in practice: MoldCo Care shows up in our provider directory and may come up in coverage. I run the AI pipeline, but MoldCo doesn't pick or shape stories. Editorial decisions are made by the pipeline based on source quality, newsworthiness, and compliance.
I do not accept referral or affiliate payments from any provider listed in the directory other than MoldCo, where I am an employee. If that ever changes, the listing will be relabeled and disclosed.
If you think I'm getting it wrong, email editorial@themoldreport.com.
The pipeline is aligned with the Shoemaker Protocol research framework. Biomarker references prioritize TGF-β1, MMP-9, MSH, C4a, and VIP. This is a known bias — I think it's the most evidence-backed framework available for mold-related illness, but mainstream medicine doesn't universally agree, and you should know that going in.
The pipeline source bias is documented in our config: peer-reviewed journals, government agencies (EPA, CDC), and accredited news organizations are weighted heavily; press releases, affiliate content, and social media are deprioritized.
If you're named in an article and want it changed or removed, email editorial@themoldreport.com with the article URL and the reason. I commit to reviewing within 5 business days. I may decline if the request is unsupported or the article is sourced to public court filings or established news.
Same address for typos, factual errors, and tip submissions: editorial@themoldreport.com.