zepto-scraper

A Python scraper for Zepto, India's quick-commerce grocery app. Two jobs:

• Category listings: start from one subcategory URL, walk its sibling subcategories and get every product card with price, MRP, discount, in-stock flag, units in stock, rating and the dark store it came from.
• Product details: turn product pages into a catalog sheet with seller entity, shelf life, FSSAI licence number, country of origin, category path, order limit and the full description.

Output as Python dicts, JSON or CSV. One dependency (requests), no browser.

pip install zepto-scraper

Why a plain request fails

Zepto sits behind AWS WAF on CloudFront:

Watch the 202. It means "Accepted", so raise_for_status() treats it as success and a scraper that only checks for errors writes empty pages without a single failure. This package treats anything but a 200 as an error.

Pages are fetched through the Scrape.do API through residential proxies in India (super=true&geoCode=in), no render: 10 credits per page. You need a Scrape.do API token; the free plan includes 1,000 credits a month (about 100 Zepto pages).

Quick start

from zepto_scraper import ZeptoScraper

scraper = ZeptoScraper(token="YOUR_SCRAPEDO_TOKEN")  # or set SCRAPEDO_TOKEN

# Any subcategory URL; the scraper walks its siblings in the same category
rows = scraper.get_category(
    "https://www.zepto.com/cn/munchies/chips-crisps/cid/d2c2a144-43cd-43e5-b308-92628fa68596/scid/df4f5100-c02f-4906-83b8-ddb744081a7a"
)
for row in rows[:3]:
    print(row["name"], row["price_inr"], row["mrp_inr"], row["in_stock"], row["available_qty"])

product = scraper.get_product(rows[0]["url"])
print(product["seller"], product["shelf_life"], product["fssai_licence"])

Start from a subcategory URL (/cn/<category>/<subcategory>/cid/.../scid/...). The category link on Zepto's homepage is not a listing.

Command line

export SCRAPEDO_TOKEN=your_token

# every product in a category, CSV
zepto-scraper category "<subcategory-url>" -f csv -o zepto-category.csv

# catalog sheet for the first 10 products from that file
zepto-scraper product -i zepto-category.csv --limit 10 -f csv -o zepto-products.csv

The two commands chain: product reads the url column of the listing CSV.

Output

Listing row (from the fictional test page in tests/fixtures):

{
  "variant_id": "44444444-4444-4444-8444-444444444440",
  "name": "Example Chilli Chips",
  "brand": "Examplo",
  "pack_size": "1 pack (41 g)",
  "price_inr": 19.0,
  "mrp_inr": 20.0,
  "discount_pct": 5,
  "in_stock": true,
  "available_qty": 30,
  "rating": 4.9,
  "rating_count": 401062,
  "subcategory": "Chips & Crisps",
  "store_id": "33333333-3333-4333-8333-333333333333",
  "url": "https://www.zepto.com/pn/example-snack-0/pvid/44444444-4444-4444-8444-444444444440"
}

Product sheet columns: variant_id, name, brand, category, pack_size, mrp_inr, max_per_order, rating, rating_count, shelf_life, country_of_origin, seller, fssai_licence, image_count, description, url.

How it works

The Next.js stream. Zepto runs on the Next.js App Router. A subcategory page makes no separate product API call; every product card is streamed into the HTML as self.__next_f.push([1, "..."]) chunks. Each chunk is a JavaScript string literal, and JSON strings use the same escaping, so json.loads('"' + chunk + '"') decodes it. Joined, the chunks are one text with one JSON object per product card.

Out-of-stock cards. Card keys are sorted alphabetically. In-stock cards start with availableQuantity; out-of-stock cards drop that key and start with baseProductId. A parser that anchors only on availableQuantity silently skips every out-of-stock product. This one anchors on both, reads each object with json.JSONDecoder.raw_decode, and dedupes on the variant id.

Prices are in paise. "discountedSellingPrice": 1900 is ₹19. The scraper divides by 100.

Siblings. Every subcategory page links its siblings, all sharing the category cid. The same subcategory can appear under a second category slug that renders no products, so only paths under the start URL's own slug are kept.

Product pages. The catalog fields come from the schema.org Product JSON-LD block (Zepto writes id before type on the script tag, so the regex does not depend on attribute order). MRP and the order limit come from the stream's productVariant object for the exact variant in the URL.

Limits

• 32 products per subcategory. Zepto's web pages stop at 32 cards (endOfPage: true) and ignore ?page=2. For wider coverage, walk more subcategories and categories.
• One dark store. Each price and stock count belongs to one dark store. With no location, Zepto serves its default store (east Bengaluru); the store_id column records which one. Changing the location through Scrape.do did not move the store in our tests.
• Use the listing for price and stock. Product pages cap the stock count and can lag the listing, so get_product() leaves price and stock out on purpose.
• No search. Search results need a browser-computed request-signature header, and robots.txt disallows /search. Walk categories instead.
• Errors. Non-200 responses (including the 202 challenge) raise ScrapeError with the status code. Server errors and 429s are retried twice; client errors are not.

Responsible use

The scraper reads category (/cn/) and product (/pn/) pages, which Zepto's robots.txt allows; it does not touch /search, /api/, cart or account paths. Product data is not personal data. Read Zepto's terms for your use case and keep request rates modest. This is not legal advice.

Development

git clone https://github.com/bugrahansss/zepto-scraper
cd zepto-scraper
pip install -e .
python -m unittest discover -s tests -v

The tests run offline against fictional fixture pages.

License

MIT

Built and tested with Scrape.do, October 2026.