Skip to content

Sephora product data by API: prices, ratings and inventory fields

Send a Sephora product URL and receive fresh product data without running proxy or browser infrastructure yourself.

OUTPUT FORMATSHTMLJSONMarkdown

One call can return the fields needed for price tracking, ingredient research, and review analysis.

What you can pull, and how to choose options

Use this guide to decide what to collect, which request options to enable, and how to check the result before scaling a run.

Decision checklist

  • Start with one product URL and confirm the fields you need: SKU, brand, price, ingredients, star rating, review count, shade variants, and image gallery.
  • If the page depends on scripts, enable JavaScript rendering; if content varies by market, add a country code; if blocking is aggressive, consider the stealth proxy option.
  • Choose selector rules when the layout is stable and you want precise field mapping; choose natural-language extraction when you want structured JSON without writing selectors.
  • Run a small batch first, inspect missing fields, then scale concurrency within your plan.

Common workflows

Three recurring jobs cover most Sephora work:

  • Price monitoring by SKU across categories, for pricing intelligence, promotion checks, and competitor merchandising.
  • Ingredient collection from product pages, for clean-beauty research, compliance datasets, and ingredient-trend analysis.
  • Review collection per product, including star ratings, counts, and review text, for review mining and merchandising.

A useful example: for a price-monitoring run, request the same SKU list on a schedule, store price and rating with a timestamp, and flag only changes. That keeps the dataset small and makes movement visible without re-parsing full pages.

How the request flow works

Create an account to obtain an API key, install the SDK for your stack or use raw requests, then send the Sephora URL with the key. The service renders the page in a real browser, rotates the IP, and returns a response your code can parse. Advanced parameters extend the basic call when a page needs them.

Parameter map and troubleshooting

render_js: run JavaScript in a headless browser and return rendered HTML. stealth_proxy: use a stealth proxy option for CAPTCHA-heavy or aggressive responses. country_code: route through a residential IP in the chosen country. ai_extract_rules: describe the fields in natural language and receive structured JSON without selectors. extract_rules: map CSS selectors to field names for deterministic parsing.

Failure distinction: a missing field after rendering usually means the extraction spec or selector is wrong; an empty or blocked response usually means routing, proxy, or rendering options need adjustment. Fix one variable at a time on a single URL before retrying a batch.

Related capabilities

Beyond product fields, the same API supports full-page screenshots, JavaScript scenarios such as clicking, scrolling and waiting, Markdown output for LLM workflows, Google search extraction for Sephora listings, and no-code automation through n8n. Pick the output format that matches the next step: JSON for databases, Markdown for analysis, screenshots for visual checks.

Output choice at a glance
OptionUse it whenReturns
Selector extraction rulesLayout is stable and fields must map preciselyParsed JSON
Natural-language extractionYou want structured output without selectorsParsed JSON
JavaScript renderingContent appears only after scripts runRendered HTML
Country routingPrice or content varies by marketMarket-specific page

Options are described from the captured product overview; exact availability depends on owner-confirmed implementation.

Frequently asked questions

Will requests be blocked?

Rotating residential proxies and a real headless browser make requests resemble ordinary browsing. For tougher responses, add the stealth option described in the parameter map below.

Is coding required?

Yes. The dashboard builder accepts a URL and parameters, then produces a ready snippet in Python, Node, Go, PHP, Ruby, or cURL.

How is JavaScript handled?

Enable JavaScript rendering so the page runs in a headless browser and the returned markup reflects the rendered state.

Can a specific country be used?

Yes. Supply a country code to route the request through an IP in that country when local content differs.

What limits apply?

Limits are set by plan concurrency rather than a per-target rate. Spacing requests or using multiple sessions helps when a target throttles aggressive patterns.

How do I get structured JSON?

Use selector-based extraction rules mapped to field names, or a natural-language extraction spec; either returns parsed JSON.