Fragrantica Scraper

Turn Fragrantica perfume pages into structured JSON. Search by keyword or paste search / perfume URLs, then collect brand, notes pyramid, main accords, ratings, images, seasonal vote breakouts, and optional review snippets.

Features

• Search by query and/or startUrls (search pages + perfume detail URLs)
• Perfume id, title, description, brand, gender, launch year (when present)
• Main accords with relative strength and color
• Notes pyramid: top / middle / base
• User rating average and vote counts
• Longevity, sillage, price-value, and seasonal vote breakouts when shown
• Primary image + extra gallery URLs
• Optional review text collection with a per-perfume cap
• omitFields to drop heavy slices (e.g. reviews, images) from the output

Input

Field	Type	Default	Description
query	string	—	Keyword search (provide query and/or startUrls)
startUrls	array	—	Search URLs and/or /perfume/...-123.html detail URLs
maxItems	integer	10	Max perfume detail pages to scrape
allReviews	boolean	false	Try to load more review content per perfume
maxItemsReviews	integer	50	Cap on review snippets per perfume
omitFields	array	—	Output keys to remove from each item
proxyConfiguration	object	—	Residential proxies recommended for production

Search example

{
  "query": "Chanel",
  "maxItems": 10,
  "allReviews": false
}

Detail URLs example

{
  "startUrls": [
    "https://www.fragrantica.com/perfume/Chanel/Chanel-No-5-Eau-de-Parfum-40069.html"
  ],
  "maxItems": 1,
  "omitFields": ["reviews", "peopleWhoLikeThisAlsoLike"]
}

Search URL + reviews

{
  "startUrls": [
    "https://www.fragrantica.com/search/?query=Avon"
  ],
  "maxItems": 5,
  "allReviews": true,
  "maxItemsReviews": 30
}

Output

One Dataset item per perfume:

Field	Description
id	Numeric perfume id from the URL
url / title / description	Page identity and blurb
brandName / brandUrl / brandLogo	Designer fields
primaryImageUrl / images	Bottle / marketing images
mainAccords	{ accord, value, color, rgb, hex }[]
pyramid	{ top, middle, base } note arrays
perfumeRating / ratingCount / reviewCount	Community scores
ratingBreakout	love / like / ok / dislike / hate counts when present
longevityBreakout / sillageBreakout / priceValueBreakout	Vote histograms
seasonBreakout	winter / spring / summer / fall / day / night
gender / year	Audience and launch year when detected
perfumers	Nose credits when listed
peopleWhoLikeThisAlsoLike	Related perfume links
reviews	Text snippets (optional)
error	Present when a single perfume failed
scrapedAt	ISO timestamp

Example shape

{
  "id": "40069",
  "title": "Chanel No 5 Eau de Parfum Chanel for women",
  "brandName": "Chanel",
  "perfumeRating": 3.57,
  "ratingCount": 12509,
  "gender": "women",
  "mainAccords": [
    { "accord": "woody", "value": 100, "hex": "#774414" },
    { "accord": "aldehydic", "value": 85.26 }
  ],
  "pyramid": {
    "top": ["Aldehydes", "Ylang-Ylang", "Neroli", "Bergamot", "Peach"],
    "middle": ["Iris", "Jasmine", "Rose", "Lily-of-the-Valley"],
    "base": ["Sandalwood", "Oakmoss", "Vanilla", "Patchouli", "Vetiver"]
  },
  "primaryImageUrl": "https://fimgs.net/mdimg/perfume-thumbs/375x500.40069.jpg"
}

Use cases

• Build fragrance catalogs and recommendation datasets
• Competitive research across designers and notes
• Enrich e-commerce listings with accords / pyramid metadata

Scope notes

• Search pages are scrolled to discover perfume links up to maxItems; very deep catalogs may need multiple queries.
• Review extraction is best-effort text snippets from the page — not a full threaded forum export.
• Vote breakout labels depend on what the page currently renders; missing sections stay empty objects / null.
• Pros / cons community lists are not always present on every perfume page.

Tips

• Start with maxItems: 3 and one brand query to validate proxies.
• Residential proxies are recommended for production.
• Use omitFields: ["reviews", "images"] for lighter rows when you only need accords and ratings.
• Prefer direct perfume URLs in startUrls when you already know the targets.