The Legal Red Lines of E-commerce Copy: Food and Supplements Especially

"Detox." "Fat-burning." "Anti-inflammatory." "Therapeutic." Put these words on a food product page and you don't have a copywriting style problem — you have a legal problem, with fines large enough to shut a small brand down outright. We run a supplement e-commerce business ourselves, and our understanding of this red line doesn't come from a compliance seminar; it comes from real regulatory pressure: the authorities patrol the internet daily, a screenshot is evidence, and consumer reports are routine. This article lays out the legal red lines for food e-commerce, and shares how we systematized content review.

What the law regulates: two acts, two lines

Food copy in Taiwan is governed mainly by two laws. Article 28 of the Act Governing Food Safety and Sanitation provides that food labeling, promotion, and advertising must not be false, exaggerated, or misleading — and must never claim medical efficacy. Violating the former carries fines of NT$40,000 to NT$4,000,000 (≈US$1,300–130,000); claiming medical efficacy carries NT$600,000 to NT$5,000,000. The Health Food Control Act adds another layer: only products certified as "health foods" (the "small green man" seal) may claim health benefits, and only in the exact approved wording — even for a genuinely certified product, extending the efficacy description on your own is still illegal, with fines in the hundreds of thousands to millions of NT dollars, imposable repeatedly per violation.

In one sentence: for ordinary food, no health-benefit claims of any kind; for certified health foods, only the approved sentences. The fine range overall runs from NT$40,000 to NT$5,000,000, and "correct within a deadline" is not a get-out-of-jail card — infringing content has to come down immediately.

The red-line vocabulary: words you cannot use

The following words, appearing in food copy, map almost directly to fines (to repeat: these are the prohibited terms — they're listed so you can audit your own copy):

• Medical-efficacy category: therapeutic effect, treat/cure, reduce inflammation, anti-inflammatory, antibacterial, repair, boost immunity, fight disease — presenting food as medicine sits in the most heavily fined tier.
• Exaggerated/misleading category: detox, fat-burning, fat loss, weight loss, slimming, speed up metabolism, flush out oil — even without mentioning any disease, claims about changing the body's appearance or physiology are violations all the same.
• The easily overlooked: meal replacement, replaces a meal, satiety, antioxidant — seemingly mild terms that, depending on context, are equally likely to be judged misleading.

The most underestimated point: the authorities judge substantive meaning, not keywords. Rewriting "fat-burning" as "feel your body turn light" is a skin-swap — reviewers read the copy in context and rule it a violation anyway. Dodging keywords doesn't work; the only safe way to write is to return to facts: ingredients, flavor, specifications, origin, and the verbatim text of any certification. And don't assume changing the medium helps — website, app, social posts, LINE group messages: anything promotional and outward-facing falls within scope.

Systematize the review: don't rely on human memory for red lines

There are dozens of red-line terms, and copy comes from product pages, campaign pages, social posts, and ad creative — relying on an editor's memory will fail eventually. Our approach is to build compliance into the system:

1. Centralized violation-term lexicon: red-line terms and exception rules live in a shared checking component, and every place that produces content uses the same copy of it.
2. Daily automated scans: a scheduled job scans all published content site-wide and alerts immediately on suspected violations — instead of finding out when someone reports you.
3. Automated pre-publish gate: new content entering version control automatically runs the compliance check; content containing prohibited terms never reaches production at all.
4. Human review at the semantic layer: because machines catch words, not meaning, high-exposure copy — rare campaign content, ad creative — always gets an extra human pass. Ads especially, since they sit outside on-site scanning and are the easiest blind spot.

Compliance is not marketing's enemy. A copy team that can't write exaggerated benefits is the one that truly learns to persuade with product facts.

If you sell food or supplements, open your product pages right now and sweep them against the list above. If you'd like this kind of automated compliance pipeline built into your own store, see our e-commerce services or talk to us directly — we run this exact machinery every day; it isn't theory.