AI visibility for beauty brands means being named when a shopper asks an AI assistant which sunscreen, serum, shampoo, lipstick or beard oil to buy. Depra AI checks four engines for it: ChatGPT, Gemini, Perplexity and Google AI Overviews (the AI summaries above some Google results). It asks your shoppers' questions from inside India, or any of 40 markets, and keeps every full answer.
Supplements and Ayurveda have their own page, Depra AI for health and wellness brands, and apparel sits under Depra AI for fashion brands. All eight industries are on the Depra AI for consumer brands hub.
How do beauty buyers find brands in AI answers?
They describe their skin or hair, a concern and often a budget, and get back a few brands and the sites the engine read. The brands can change between runs, engines and languages.
Depra AI's study asked five skincare questions from India. Each named a skin type or a concern, such as oily skin, dark spots or a white cast. Two set a rupee price cap, one named vitamin C and one Indian winters (study prompt file, checked 8 Oct 2026).
Half the questions in Depra AI's study of 480 AI shopping answers were about skincare, asked of ChatGPT, Gemini and Perplexity in India on 14 August 2026. Of its 4,640 citations, youtube.com drew 534, reddit.com 191 and nykaa.com, a beauty retailer, 176.
By answers, reddit.com was cited in 106 of 480 AI shopping answers and youtube.com in 100, the top two on that count. (Hinglish vs English AI shopping study, India, 14 Aug 2026)
AI visibility for beauty brands: what does Depra AI record, from sunscreen to beard oil?
It records what four engines say to your shoppers' questions, checking ChatGPT, Gemini and Google AI Overviews each day and Perplexity each week.
- Mentions and rank: how often "sunscreen for oily skin" answers name you (visibility), your place in the list, tone, and share of all brand mentions (share of voice).
- One product, one question: for "best niacinamide serum under ₹600", whether each answer names your serum, a rival's or neither, on all four engines.
- Products ChatGPT shows: on every plan, the Shopping report ranks brands by their share of answers with product cards, and lists the stores they are sold by. Depra AI records product cards from ChatGPT only. OpenAI says shopping in ChatGPT is live for users in the US today and will expand to more regions (OpenAI merchants page, checked 7 Oct 2026). Questions tracked for India may show few or no cards until ChatGPT shows products there.
- Sources and gaps: each cited site labelled by type, such as marketplace or forum. On Growth and above, the sites cited for a rival but not you, such as a "best moisturiser" list.
- Perception and facts: qualities answers give each brand, such as "non-greasy", and claims backing or contradicting your facts, such as a sunscreen's price, over 28 days.
- Changes and actions: what changed between your two latest scan days, plus a ranked list of pages to publish, sites to get cited on and own-site checks.
- Languages and markets: English and Hinglish (Hindi and English typed in English letters); a project can cover up to 10 of the 40 markets.
How do I set up Depra AI for a beauty or grooming brand?
Yourself, in five steps: Depra AI is self-serve. The Depra AI quickstart has the detail.
- 1. Create an account and give your website. Setup runs a first free scan of ChatGPT, Gemini and Perplexity, with no card needed.
- 2. Check the brand profile Depra AI drafts against your real ranges, such as sunscreens or beard care, and your price band.
- 3. Add markets and rivals: India plus any country you ship to, and up to 3, 5, 8 or 12 competing brands, by plan.
- 4. Pick your questions. Keep the suggested ones (up to 5 topics of 8, none naming you) that fit your skin types, hair concerns and price points.
- 5. Start the trial. Choose Starter, Growth or Scale and add a card or UPI AutoPay: 3 days free, first charge on day 4 unless you cancel. Then tracking begins.
What does Depra AI read from my beauty store, and how?
DepraBot, the web reader behind Depra AI, reads the product list your store makes public, so Depra AI knows which serums and shampoos are yours. It never signs in, and Depra AI never changes your store.
For a Shopify store the source is the Storefront API at your myshopify.com address, read with no access token, plus /products.json if that list is incomplete. Other stores go through the WooCommerce Store API or sitemaps.
Nothing is read beyond your site, its subdomains and your myshopify.com address. DepraBot follows robots.txt, the file where a site tells bots what they may read; if yours shuts DepraBot out, upload your product list as a file. The list is scanned at setup, weekly and when you press Fetch. Fetch allows 3 scans a day per project on Starter and Growth, and 5 on Scale and Enterprise.
Questions are asked from inside the country you pick, and answers are stored in full. About DepraBot and the store catalogue docs give the full rules.
A Depra AI app for Shopify is coming to the Shopify App Store. When it is listed, it will read your products, collections and store pages. Until then, setup needs no Shopify app and no access token.
How do I connect Depra AI to Claude and other AI apps?
In six steps, through the read-only Depra AI MCP server. MCP (Model Context Protocol) lets Claude and similar apps query your stored Depra AI data.
- 1. Go to API Keys, listed under Company in the Depra AI app's sidebar.
- 2. Press Create token to make a personal access token.
- 3. Open the tab for your app: Claude Code, Cursor, VS Code, Claude Desktop, Codex or Other.
- 4. Follow that tab to add the server address and your token.
- 5. Do what the tab asks before testing: quit and reopen Claude Desktop, or open a new terminal for Claude Code or Codex. In VS Code, start the server and paste your token.
- 6. Ask a test question: "List my Depra AI projects."
What can the app see?
Only stored data: the server cannot change a project, start a scan or edit your site. Keep your token secret, because a token can read only the projects its creator can access. ChatGPT and claude.ai on the web cannot connect yet. MCP setup has the exact files; see also how the server works and all 24 tools.
Which buyer prompts should a beauty brand track first?
Start with the details beauty shoppers give: skin type, hair concern, climate, shade and budget. Track one or two per hero product. These 12 are examples to adapt, not results. None names a brand, because branded questions are left out of your scores by default.
A slot is one question in one language and one market, and Starter has 16. Prompts and topics covers slots, topics and tags.
| Prompt | What it shows you | Type |
|---|---|---|
| Which SPF 50 sunscreen suits oily skin in humid weather? | Whether answers name your sunscreen when a buyer pairs a skin type with the climate | Problem, climate |
| Best niacinamide serum under ₹600 for oily skin | Which brands are named for one ingredient inside one price band | Ingredient, price band |
| Niacinamide or vitamin C serum: which is better for dull skin? | Which serums answers pick when a shopper weighs two ingredients | Comparison |
| Best shampoo for hair fall from hard water | Whether your haircare range is named for a concern tied to water at home | Problem |
| Which moisturiser works for combination skin during the monsoon? | Whether seasonal questions name you, and which sources they cite | Climate |
| Best long-lasting matte lipstick under ₹800 | Which makeup brands, and on ChatGPT which products, appear for a price-capped question | Price band, product |
| Which kajal does not smudge in summer? | Whether one hero product is named against rival products | Product |
| Tinted sunscreen that suits deeper Indian skin tones | Whether answers name you for your shade range, and which sources they cite | Skin tone |
| Best perfume gift set for men under ₹2,500 for Diwali | Whether fragrance gifting questions name you before the festival | Occasion, price band |
| Best beard oil and trimmer kit for a first beard | Whether men's grooming answers name your kit or a rival's | Men's grooming, product |
| Fragrance-free body lotion for sensitive skin | Whether answers repeat what your label says, such as fragrance-free | Problem, ingredient |
| Skincare routine to start three months before my wedding | Whether routine questions, not only product questions, name your products | Occasion |
Which SEO, AEO and GEO audit prompts can I ask Claude about my beauty brand's Depra AI data?
Paste any row into a connected app. It answers from your stored Depra AI data with the read-only tools in the third column. MCP recipes has more.
Rows 1 to 4 are GEO (generative engine optimization): whether answers name you, against rivals. Rows 5 to 8 are AEO (answer engine optimization): the claims answers make about you and the questions your pages should cover. Rows 9 and 10 are SEO: the pages and domains AI cites. No MCP tool checks crawlers, so test robots.txt in the app's AI crawler checker.
| Prompt | What it checks | Depra AI MCP tools it uses |
|---|---|---|
| Across our sunscreen questions this month, what share of answers on each engine name us, and where do we sit in the list against our closest rivals? | GEO: how often and how high each engine names you on sunscreen questions | list_projects, list_topics, get_brand_report |
| In our acne-care answers, which rival comes before us most often when both are named? Open the two newest answers where that happened. | GEO: who is named first when you and one rival both appear | list_brands, get_brand_insights, get_head_to_head_chats, get_chat |
| Is our haircare topic weaker in Hinglish than in English? Show the gap for each engine. | GEO: the English and Hinglish gap in one product line | get_brand_insights, get_brand_report |
| Which lipstick and kajal brands did ChatGPT show as product cards, at what average position, and which stores were they sold by? | GEO: product cards in ChatGPT answers, with the top merchants | get_shopping_summary, list_chats |
| Do any AI answers get our facts wrong, such as our sunscreen's SPF, price or pack size? Quote two of them. | AEO: claims in answers that contradict your facts | list_facts, list_chats, get_chat |
| Which words do answers use for us and for rivals in skincare, such as "gentle" or "greasy", and where is our tone most negative? | AEO: brand perception and the tone of mentions | get_brand_perception, get_sentiment_report |
| Which web searches did ChatGPT run behind our niacinamide and vitamin C prompts? Turn them into questions our serum pages should answer. | AEO: the questions your product pages should answer | list_search_queries, list_prompts |
| Which open actions target our fragrance gifting topic? Walk me through the evidence and steps of the first. | AEO: what to do next, and why | list_actions, get_action |
| For our serum topic, which domains and pages are cited most, and which of them are marketplaces, forums or videos? | SEO: the domains and pages AI cites, by type | get_domain_report, get_url_report |
| For hair fall questions, list the pages cited when a rival is named and we are not, without low-confidence rows. What does the top page say? | SEO: page-level gaps against rivals | get_source_gaps, get_url_content |
What helps a skincare or makeup brand get recommended by AI?
Back every claim, publish ingredients as text, complete your listings and earn honest mentions. None of this guarantees a recommendation.
Make every claim one you can back up
Rule 36 of India's Cosmetics Rules, 2020 bars any cosmetic from claiming or conveying a false or misleading idea (CDSCO, checked 8 Oct 2026). The code of the Advertising Standards Council of India (ASCI) says factual claims should be capable of substantiation (ASCI Code, checked 8 Oct 2026).
ASCI's guidelines for skin lightening or fairness products add that advertising should not communicate any discrimination as a result of skin colour (ASCI guidelines, checked 8 Oct 2026). Add your key claims as facts in Depra AI to see how often answers back or contradict them.
Publish your full ingredient list as text
The Cosmetics Rules, 2020 require labels to list ingredients under the word INGREDIENTS, with those above one percent first, in descending order. Packs of up to 60 ml of liquid or 30 g of solid or semi-solid are exempt (rule 34(7), Cosmetics Rules, 2020, checked 8 Oct 2026). Put the same list as text on your product pages and listings.
Answer questions by skin type, concern and weather
Write pages that say which of your products is meant for oily skin in humid weather, dry skin in winter or deeper skin tones. Take the exact phrases from ChatGPT's own web searches behind your answers, which Depra AI stores.
Complete your listings on beauty stores and marketplaces
Nykaa lists 2,400+ brands across makeup, skincare, haircare, fragrance and personal care (Nykaa, checked 8 Oct 2026). Purplle sells beauty and cosmetic products online (Purplle, checked 8 Oct 2026), and so does Flipkart (Flipkart, checked 8 Oct 2026). Reliance Retail launched its Tira beauty platform on 5 April 2023 (Reliance Retail, checked 8 Oct 2026).
nykaa.com drew 176 citations in Depra AI's study. Where answers cite a store with no page for you, create or claim your listing so the page exists for an engine to find and cite. See the marketplace visibility guide.
Work with creators, and label every paid post
In Depra AI's study, youtube.com drew more citations than any other site, so creator videos are pages answers can cite. ASCI's guidelines require a disclosure label for any material connection with the brand, free products included, placed upfront and prominently (ASCI influencer guidelines, checked 8 Oct 2026). Influencers must also check that brands can substantiate claims, so give creators only claims you can prove.
Take part in forum threads as the brand, openly
reddit.com appeared in more of the study's answers than any other site. Reply to skincare and haircare threads under your brand's own name, and never pay for hidden promotion.
What does Depra AI not do for beauty brands?
Depra AI only measures AI answers; your store, listings and content stay with you.
- No marketplace data: no rankings inside Nykaa, Amazon or Flipkart search, and no marketplace sales, prices, stock, ads or reviews.
- Product cards from ChatGPT only: for a lipstick question, only ChatGPT's product cards are recorded; the other engines count for mentions and sources.
- No copy or store edits: it does not write product descriptions or blogs, edit your site or change your store.
- No compliance review: it does not check claims, ads or labels against the ASCI Code or the Cosmetics Rules.
- No guarantees: the before-and-after view of a finished action is not proof of cause, and no mention is promised.