Carousell Scraper - Extract Listing Data from Carousell

Carousell Listing Fields and Price Data We Extract

A Carousell listing is a product record first: the page carries schema.org Product markup with price, currency and item condition, repeated in a Details block. Every category shares a core:

• Listing id and URL - the address is /p/ plus a title slug ending in a numeric id, and that id is the only stable key: the slug follows a title the seller can rewrite.
• Title, description, price and formatted price - raw amount and display string with its currency prefix arrive as separate values.
• Condition - Brand new, Like new, Lightly used, Well used or Used, shipped as a numeric code behind the label.
• Category, breadcrumb and posted date - leaf name with its numeric id, and relative text such as 27 minutes ago resolved against collection time.
• Likes and offers - the like counter sits on the card and the listing, the offer count records how many buyers bargained, and a Make Offer control sits on each one.
• Photos and deal method - the image list on media.karousell.com with a photo count, and meet-up at a named MRT station or mailing.
• Seller - username, join date, rating, split of positive, neutral and negative reviews, live and sold counts, and badges: Professional Seller, Preferred, Verified Landlord, Certified Partner.
• State - live, Reserved or Sold, set by the seller rather than inferred.

Cars are a marketplace inside the marketplace: mileage in km, first registration date, annual depreciation, a monthly installment, and a Financials block with OMV, ARF and COE expiry, beside propellant, transmission, engine capacity in cc, max power output in bhp, manufactured year, doors, seats and owner count.

Property rentals carry Furnishing, Lease, Level, Type, Room and Gender, a features grid of Air-Conditioning, Cooking, Internet, No Owner Staying, PUB Included and Visitors allowed, plus postal code, street name and nearest MRT stations, and price per square foot beside the rent. Fashion and Luxury add Brand, Model, Type, Gender and a size given in EU, UK and US at once. Carousell Certified items sit on catalogue pages carrying a product id and a variant id, with battery health, warranty flag, box contents and the inspecting partner and date.

Carousell Data Beyond the Listing Body

Beyond the listing body, what decides a Carousell scraping plan is how the site is addressed and where it stops answering:

• Address grammar - a category path stacks slugs that each end in a numeric id, a keyword search is a slug ending in /q/, a search scoped to one category ends in /q- plus the category id, hub pages end in /h- plus an id, and Cars and Property add landing pages by make, model and estate.
• Facets in the path - price bands and brand are extra path segments, so a facet combination has its own crawlable address, while every query-string variant is closed to crawlers.
• The counter is not a total - browse and search print a rounded figure with a plus sign, then load more by an opaque continuation token instead of page numbers. Coverage comes from slicing a category by subcategory, price band, location and sort order, not from paging one query to the end.
• Sitemaps - each site names one sitemap index holding category files, search-landing files, make and model files, and per-category listing files whose URLs carry a lastmod timestamp, the cheapest way to spot new stock.
• Listing lifecycle - Property, Autos, Jobs and Services listings run on a validity window, then drop into an inactive state buyers never see, and free quota there is tight. Stock churns, so refresh interval is part of the brief.

What is absent matters as much: Carousell shows no public view counter. View and hourly insight figures live in the seller CarouBiz dashboard, so they are not in what we deliver.