cd /news/ai-tools/tracking-sephora-prices-sales-and-ne… · home › topics › ai-tools › article
[ARTICLE · art-172749] src=dev.to ↗ pub= topic=ai-tools verified=true sentiment=↑ positive

Tracking Sephora prices, sales and new launches with Python (no browser, no API key)

An 18-year-old engineering student built the Sephora Products Scraper, a paid Apify actor that turns keyword searches, categories, brand pages and product URLs into clean JSON rows over plain HTTP, without a headless browser or API key. The tool reads Sephora's JSON catalog endpoint and the product records embedded in product pages, paces requests and rotates proxy sessions to get past Akamai bot protection, and supports filters for price, rating, brand, sale and new items across the US and Canadian stores. The author also demonstrates a daily sale tracker built on the actor's output.

key takeaway

The Sephora Products Scraper returns typed numeric prices plus per-variant price, size and stock flags by reading Sephora's JSON catalog endpoint and embedded product records over plain HTTP, avoiding the blocked headless-browser route.

by read5 min views2 publishedNov 19, 2026

I built this actor; it's a paid tool on Apify with a free trial credit. This article was drafted with AI and published by me.

Sephora is where beauty brands launch, so its catalog answers a lot of practical questions: what does a brand actually sell there and at what prices, which products are on sale this week, what launched in the last month, which face serums are rated 4 stars and up. The site has no export and no public API, and the obvious route (search page HTML in a headless browser) is slow and gets blocked.

I'm an 18-year-old engineering student, and I built Sephora Products Scraper to turn keyword searches, categories, brand pages and product URLs into clean JSON rows over plain HTTP. This post covers how it works, calling it from Python and Node, what a row looks like, and a small project: a daily sale tracker that tells you what went on sale and by how much.

Sephora's website loads its listings from a JSON catalog endpoint (about 100 products per request) and renders each product page with the full product record embedded in the HTML: every shade or size with its own price and stock flag, the ingredient list, the description, loves count and the review sentiment tags. The actor reads exactly those two things, paces requests and rotates proxy sessions because Sephora's Akamai protection blocks IPs that send bursts, and never touches the HTML search and browse pages that Sephora's robots.txt disallows.

Two modes:

variants array with per-shade price, size, isOutOfStock and isOnlyFewLeft. Filters and sort are applied by Sephora before anything is charged: price range (minPrice, maxPrice), minimum rating ( minRating), only these brand names ( brandFilter), on sale only ( onSaleOnly), new only ( newOnly); sort by best selling, top rated, new or price. Both the US store (USD) and Sephora Canada (CAD) work with country: "us" | "ca".

from apify_client import ApifyClient

client = ApifyClient("<YOUR_APIFY_TOKEN>")

run = client.actor("rel8ble/sephora-products-scraper").call(run_input={
    "categories": ["face-serum"],
    "sortBy": "bestSelling",
    "minRating": "4",
    "maxItemsPerQuery": 50,
})

for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["brand"], "|", item["name"], "|", item["price"], item["currency"], "|", item["rating"], f"({item['reviews']} reviews)")

The same in Node:

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: '<YOUR_APIFY_TOKEN>' });
const run = await client.actor('rel8ble/sephora-products-scraper').call({
    brands: ['the-ordinary'],
    country: 'ca',
    includeDetails: true,
    maxItemsPerQuery: 0,
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
for (const p of items) console.log(p.name, p.price, p.currency, p.variants.map((v) => `${v.name}: ${v.isOutOfStock ? 'OUT' : 'in stock'}`).join(', '));

A listing row from a mascara search (22 Oct 2026):

{
    "productId": "P467208",
    "name": "Lash Idôle Lengthening & Volumizing Mascara",
    "brand": "Lancôme",
    "listPrice": "$16.00 - $30.00",
    "price": 16,
    "priceMax": 30,
    "salePrice": null,
    "currency": "USD",
    "rating": 4.4998,
    "reviews": 10785,
    "variantsCount": 2,
    "isBestseller": true,
    "isNew": false,
    "url": "https://www.sephora.com/product/lancome-lash-idole-lash-lifting-volumizing-mascara-P467208?skuId=2417145",
    "position": 1,
    "sourceType": "search",
    "sourceQuery": "mascara"
}

With details on, the same product also carries description, ingredients, lovesCount, category (Makeup > Eye > Mascara), images, reviewSummary, sentiments ({"name": "Applicator", "type": "pros", "count": 76} ...) and variants, one entry per shade or size with its own price, size, isOutOfStock and url.

Prices are numbers, not strings, and regularPrice, salePrice and an isOnSale flag are separate fields, which is what makes the tracker below a 40-line script.

The idea: run the actor once a day on the categories you care about with onSaleOnly: true, keep yesterday's rows in a CSV, and print what is newly on sale, what got cheaper, and what dropped off sale. Sephora's sale prices change on a schedule (weekly events, brand promotions), so a daily diff catches almost everything.

import csv, os, datetime
from apify_client import ApifyClient

CATEGORIES = ["face-serum", "moisturizer-skincare", "mascara", "lipstick", "fragrance-women"]
client = ApifyClient("<YOUR_APIFY_TOKEN>")

run = client.actor("rel8ble/sephora-products-scraper").call(run_input={
    "categories": CATEGORIES,
    "onSaleOnly": True,
    "maxItemsPerQuery": 0,
})
today = {}
for it in client.dataset(run["defaultDatasetId"]).iterate_items():
    today[it["productId"]] = {
        "brand": it["brand"], "name": it["name"],
        "regular": it["regularPrice"], "sale": it["salePrice"], "url": it["url"],
    }

path = "sephora-sales.csv"
yesterday = {}
if os.path.exists(path):
    with open(path, newline="", encoding="utf-8") as f:
        for row in csv.DictReader(f):
            yesterday[row["productId"]] = row

new = [p for p in today if p not in yesterday]
cheaper = [p for p in today if p in yesterday and float(yesterday[p]["sale"]) > today[p]["sale"]]
ended = [p for p in yesterday if p not in today]

print(f"{datetime.date.today()}: {len(today)} products on sale, {len(new)} new, {len(cheaper)} cheaper, {len(ended)} ended")
for p in sorted(new, key=lambda p: today[p]["regular"] - today[p]["sale"], reverse=True)[:15]:
    t = today[p]
    off = round(100 * (1 - t["sale"] / t["regular"]))
    print(f"  NEW  {t['brand']} - {t['name']}: {t['regular']} -> {t['sale']} ({off}% off)  {t['url']}")
for p in cheaper:
    t = today[p]
    print(f"  DOWN {t['brand']} - {t['name']}: {yesterday[p]['sale']} -> {t['sale']}")

with open(path, "w", newline="", encoding="utf-8") as f:
    w = csv.DictWriter(f, fieldnames=["productId", "brand", "name", "regular", "sale", "url"])
    w.writeheader()
    for p, t in today.items():
        w.writerow({"productId": p, **t})

Cost scales with rows, not categories: at $0.85 per 1,000 products, 100 sale items a day is $0.085 and a full 1,000 is $0.85. Schedule the actor on Apify (or the script with cron) and pipe the output into Slack or a Google Sheet through Make, Zapier or n8n and you have a daily sale alert.

Three variations that are one input change away:

newOnly: true, sortBy: "newest" instead of onSaleOnly. brands: ["rare-beauty-by-selena-gomez"], includeDetails: true, maxItemsPerQuery: 0 gives every product, every shade and its stock flag; diff ingredients locally for Niacinamide or Retinol. $0.85 per 1,000 products, no start fee, same price with or without details. Failed requests are never charged. A keyword search returns what Sephora's search returns (a nonsense keyword gets Sephora's fallback picks rather than nothing), so use categories or brands when you need an exact set. Details cost one request per product (about 1 MB each); a listing-only run does about 100 products per request. Reviews themselves are not collected, only rating, review count, loves and sentiment tags. US and Canada only; Sephora's other sites (sephora.fr, .co.uk, .com.au) are separate platforms and not covered.

Store page: https://apify.com/rel8ble/sephora-products-scraper

Need a custom scraper or a ready dataset? Hire me on Fiverr: https://www.fiverr.com/rel8ble

── more in #ai-tools 4 stories · sorted by recency
── more on @sephora 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/tracking-sephora-pri…] indexed:0 read:5min 2026-11-19 · —