ChatGPT Ads for Supplements and Wellness: Staying Compliant
Supplements and wellness are restricted on ChatGPT ads, not banned. Here is how to write copy that passes review, which claims to avoid and how to structure a test.
Supplement and wellness brands face a familiar problem on every ad platform: the claims that sell are often the claims you are not allowed to make. ChatGPT ads are no different, and arguably stricter. Health, supplements and wellness sit in OpenAI's restricted category, and the platform's general rules against unverifiable claims, guaranteed outcomes and exploiting insecurity hit this category harder than most.
Restricted does not mean impossible. It means more review, more care and copy that describes rather than promises. This guide is about getting supplement and wellness ads approved and running in a way that holds up.
This article is general marketing guidance, not legal or regulatory advice. Supplement claims are also governed by laws and regulators in each market, such as the FTC and FDA in the US. Check with a qualified advisor before making health-related claims.
Where supplements sit in ChatGPT's ad policies
As publicly summarized in 2026, OpenAI's ad policies treat these categories as restricted: health, supplements and wellness, financial products, alcohol, dating, weight management and products for minors. Restricted ads face extra scrutiny and tighter rules about what can be said.
Several general prohibitions also apply directly:
- Deceptive or unverifiable claims.
- Invented statistics and fake reviews.
- Guaranteed outcomes, explicitly including health cures.
- Before/after transformations.
- Exploiting hardship or insecurity, or implying the ad knows what the user discussed.
- Prohibited products: drugs, and anything sold as a drug substitute, are out entirely.
Weight management is its own restricted category, so a "fat burner" or appetite product carries double the risk. Products intended for minors add another layer.
What you can and cannot say
The safest copy describes what is in the product, how it is made and how it is meant to be used. It avoids claiming that it treats, cures, prevents or reverses anything.
| Avoid | Why | Safer alternative |
|---|---|---|
| "Cures anxiety naturally" | Disease claim, health cure | "Magnesium glycinate and L-theanine in one capsule" |
| "Lose 10 lbs without dieting" | Guaranteed outcome, weight management | Usually avoid weight claims entirely |
| "Boosts immunity by 300%" | Invented or unverifiable statistic | "Vitamin D3 and zinc, third-party tested" |
| "Doctors hate this" | Clickbait, deceptive | Do not use |
| "Finally sleep through the night" | Implied guaranteed outcome, insecurity | "A non-habit-forming evening supplement routine" (only if accurate and supported) |
| "Struggling with low energy?" | Plays on a problem the user may have mentioned | "B-complex vitamins in a once-daily tablet" |
Two principles cover most cases. First, say what the product is, not what it will do to the person. Second, every factual claim needs evidence on the landing page: the dosage, the third-party test certificate, the ingredient source.
Before submitting, run drafts through the ChatGPT ad policy checker. For a broader look at claim substantiation across platforms, see ad claims compliance for marketers.
Example chat cards that describe, not promise
Headline up to 50 characters, body up to about 100.
Magnesium supplement
- Headline: Magnesium Glycinate in a Daily Capsule
- Body: 200mg per serving, third-party tested for purity. No artificial colors. Free US shipping.
Protein powder
- Headline: Unflavored Whey Protein, One Ingredient
- Body: Grass-fed whey isolate with nothing added. 25g protein per scoop. 30-day return policy.
Greens powder
- Headline: A Greens Powder Without Added Sugar
- Body: 12 vegetable and fruit powders, unsweetened. Mixes into water or smoothies. Full label online.
Sleep routine product
- Headline: An Evening Tea Blend With Chamomile
- Body: Caffeine-free chamomile, lemon balm and lavender. Loose leaf or tea bags. Ships in 2 days.
Notice the pattern: ingredient, amount, format, testing and logistics. No benefits stated as outcomes. That may feel weaker than Meta copy you are used to, but ChatGPT users asked a question already. Often they want to know exactly what is in the product.
Context hints for wellness
Context hints are not shown to users, but they still shape where the ad appears, and they should not target vulnerable situations. Describe a shopper's interest and intent, not a medical condition:
- "Person comparing forms of magnesium supplements and wanting to know which type to buy"
- "Gym-goer looking for a plain protein powder with no sweeteners or additives"
- "Someone who wants an easy way to add more vegetables to a busy routine"
Avoid hints like "person with anxiety disorder" or "someone with insomnia who can't sleep". Targeting medical conditions is exactly the type of exploitation the policy is designed to stop, and it risks your account. The context hints generator is a good starting point; review its output with this in mind.
Landing pages under extra scrutiny
For restricted categories, the landing page matters as much as the ad. The general rules in our guide to landing pages that convert ChatGPT ad traffic still apply, but reviewers and users will also look for:
- The exact claims from the card, with nothing more aggressive further down the page.
- A full ingredient list and supplement facts panel.
- Third-party testing certificates or links to batch results, if you claim testing.
- Required disclaimers for your market.
- Clear pricing, subscription terms and how to cancel. Hidden auto-renewals are a fast route to complaints.
- No before/after photos, unverified testimonials or "miracle" language anywhere on the page.
Check your ad and page together with the message match checker. If your sales page was written for an aggressive Meta funnel, build a calmer version for ChatGPT traffic.
The economics of a supplement test
Supplements usually run on subscriptions and repeat orders, so first-order math rarely works alone.
Say you sell a $39 monthly magnesium supply with a 65% gross margin:
- Gross profit per order: $39 × 0.65 = $25.35
- CPC: $4 (OpenAI's recommended starting bids have generally landed between $3 and $5)
- Landing page conversion rate: 3.5%
- Cost per new customer: $4 / 0.035 = $114.29
- Average subscriber stays 5 months: lifetime gross profit = $25.35 × 5 = $126.75
That gives an LTV to CAC ratio of about 1.11 on gross profit, which is thin. Raising conversion to 5% drops cost per customer to $80 and lifts the ratio to about 1.58. A 3-month starter bundle at checkout could raise first-order value further. Use the LTV to CAC calculator with your real retention numbers, not hopeful ones.
Because restricted ads may take longer to approve, build a week of buffer into your launch plan.
A compliant launch checklist
- List every claim in your planned ads and confirm you can document each one.
- Rewrite outcome claims as product descriptions.
- Audit the landing page for anything that would fail review, including old testimonials.
- Write context hints around shopping intent, not medical conditions.
- Install the OpenAI pixel and Conversions API, with subscription events.
- Launch one campaign at $50 to $100 a day, on CPC, with three cards.
- Allow two to three weeks before judging, and watch for disapprovals daily in week one.
How SecondWin helps supplement and wellness brands
SecondWin writes and manages ChatGPT ads for you, and policy checking is built into the process rather than added at the end. We read your site, find the longest-running ads in your niche in Meta's Ad Library, and extract the message behind them. Then we rewrite that message for ChatGPT under strict rules: no unproven claims, no health outcomes, no invented facts and no competitor names. Every ad is checked against OpenAI's policies before it runs in your own OpenAI ad account. If a product or claim cannot be advertised compliantly, we will say so.
Get a free analysis of your wellness brand, or see how plans compare on our pricing page.
FAQ
Are supplements allowed in ChatGPT ads?
Supplements, health and wellness products are a restricted category, not a prohibited one, as of OpenAI's publicly summarized 2026 policies. That means extra review and tighter rules. Drugs and products positioned as drug alternatives are prohibited. Keep copy to factual product descriptions and avoid any claim that a product treats, cures or prevents a condition.
Can I make health benefit claims in a ChatGPT ad?
Health outcome claims are where most supplement ads fail. Guaranteed outcomes and cures are prohibited, and unverifiable claims are not allowed. Even accurate, substantiated benefits face extra scrutiny in a restricted category. Describing ingredients, amounts, testing and format is the most reliable route to approval, alongside any disclaimers required in your market.
Can I target people with specific health conditions?
You should not. ChatGPT ad policies prohibit exploiting hardship or insecurity and implying the ad knows what the user discussed. Write context hints around shopping intent, such as comparing supplement forms or looking for additive-free products, rather than medical diagnoses. This protects users and your ad account.
How long does approval take for restricted categories?
OpenAI does not publish a fixed review time, and restricted categories can take longer than standard ads. Plan for a buffer of several days before your intended launch, submit early, and watch for disapprovals during the first week. A clean landing page that matches the ad reduces back-and-forth.