Coles AU Scraper (Products, Prices, Specials)

What is Coles AU Scraper?

Coles AU Scraper collects product search results from coles.com.au, one of Australia's two largest supermarket chains. Give it a keyword such as milk or oat milk and it returns one row per product: name, brand, pack size, price in AUD, was-price on specials, unit price, stock availability, category, image, product link and product id. It reads the search data that the Coles search page embeds in its HTML, so no login is needed.

What data does Coles AU Scraper extract?

Field	Type	Description
sku	string	Coles product id (stable key).
name	string	Short product name, e.g. Full Cream Milk.
brand	string	Brand as listed, e.g. Coles, Devondale.
fullName	string	Long product name with brand and size.
packageSize	string	Pack size, e.g. 3L.
price	number	Current price in AUD.
wasPrice	number	Previous price when it is above price; otherwise null.
onSpecial	boolean	Online special flag from the price block.
unitPrice	number	Price per unit of measure.
unitPriceText	string	Unit price as shown, e.g. $1.65/ 1L.
unitOfMeasure	string	Unit of the unit price, e.g. 1 l.
inStock	boolean	Availability flag from the search data.
availabilityType	string	e.g. InStoreAndOnline.
category, subCategory, aisle	string	Coles category tree of the first online heir.
imageUrl	string	Product image from the Coles image host.
url	string	Product page on coles.com.au (/product/<name-and-size>-<id>).
isSponsoredAd	boolean	True for sponsored product tiles.
searchTerm, page, position	mixed	Keyword, search page and rank.
totalResults	number	Total hits Coles reported for the keyword.
scrapedAt	string	ISO time of the scrape.

Banner tiles that Coles mixes into search results are skipped; only product tiles become rows.

How to use Coles AU Scraper

1. Open the Actor and enter one or more keywords separated by ;, for example milk; bread; eggs.
2. Set Max products per search (default 100). Each keyword is scraped separately, up to this limit. About 48 products come per page.
3. Leave Proxy mode on auto. The Actor tries a direct connection first and switches to Apify's Australian residential proxy only when Coles blocks the request.
4. Click Start, then download the dataset as JSON, CSV or Excel, or read it through the API.

Example input

{
    "keywords": "milk; bread",
    "maxProducts": 60
}

Example output

Real row from a run on 2026-10-23 (keyword milk, sixth result, sponsored tile):

[
  {
    "sku": "1329781",
    "name": "High Protein Full Cream Milk",
    "brand": "Devondale",
    "fullName": "DEVONDALE HIGH PROTEIN FULL CREAM MILK 1L",
    "packageSize": "1L",
    "price": 2,
    "wasPrice": 2.5,
    "onSpecial": true,
    "unitPrice": 2,
    "unitPriceText": "$2.00/ 1L",
    "unitOfMeasure": "1 l",
    "inStock": true,
    "availabilityType": "InStoreAndOnline",
    "category": "Long-Life Milk",
    "subCategory": "Drinks",
    "aisle": "Long Life Milk",
    "imageUrl": "https://productimages.coles.com.au/productimages/1/1329781.jpg",
    "url": "https://www.coles.com.au/product/high-protein-full-cream-milk-1l-1329781",
    "isSponsoredAd": true,
    "searchTerm": "milk",
    "page": 1,
    "position": 6,
    "totalResults": 323
  }
]

Input parameters

Parameter	Default	Meaning
keywords	milk	Search terms separated by ;.
maxProducts	100	Products per keyword, 1 to 5000.
proxyMode	auto	auto (direct, residential only after a block), never (direct only), always (residential from the start).

Pricing

Pay per result: $0.80 per 1,000 products on the free tier, falling to $0.60 per 1,000 on the highest tiers. The Actor start is charged separately ($0.05 per 1,000 starts). Measured on 2026-10-23: a 120-product run cost $0.0016 at 1 GB, direct only. A 639-product run (4 keywords) cost $0.0165 at 1 GB, including 1.07 MB of residential proxy traffic ($0.008) after Coles blocked one direct page request; the Actor switched to residential and the search continued. The closest public Coles scraper, dromb/coles-au-product-search-specials-stores-unofficial (278 monthly users on 2026-10-23), charges $1.00 per 1,000 results on its free tier and $0.70 at its top tiers.

Coles AU Scraper vs. the alternative

The closest public Coles scraper charges $1.00 per 1,000 results on its free tier. This Actor charges $0.80 on the same tier and returns the unit price, was-price, category and product id in every row.

Using Coles AU Scraper with AI agents / MCP

The Actor is pay per result and takes a plain keyword input, so an AI agent can call it through the Apify MCP server with {"keywords": "milk", "maxProducts": 50}. The output is one flat row per product.

FAQ

Does it need a Coles account? No. It reads the public search page.

Why do some rows have a was-price? Coles shows a previous price on many products. In the 120 test rows, 58 had a was-price above the current price. onSpecial is the separate online special flag from the same price block.

Are prices per store? No. The search page shows the online price set. Store-level prices are not covered.

Is the url and imageUrl verified? Yes. The sample product page and image both returned HTTP 200 when checked on 2026-10-23.

What happens when Coles blocks the Actor? Imperva sometimes answers with a page that has no search data. The Actor retries after short waits and then switches to Apify's Australian residential proxy. Searches that still fail go into the ERRORS record and are not charged.

Can I scrape a category page? Not yet. The Actor searches by keyword only.

What if the same product matches two keywords? It is returned once, under the first keyword that found it, and counted once in the charge.

Limitations

• Keyword search only. No category browsing and no store-level prices or stock.
• inStock is the search-data availability flag, not a live stock count.
• A search returns at most the hits Coles reports (totalResults); the Actor stops at maxProducts.
• Scale tested: 639 products in one run. One direct page was blocked and fetched through the residential proxy. Heavy, sustained blocking was not tested.

Support

Open an issue on the Actor's Issues tab. Include the keyword, the input JSON and the run ID.