Flipkart Product Sellers Scraper

Collect public seller offers from Flipkart’s All Sellers comparison pages. The Actor accepts product URLs, direct sellers?pid= URLs, or a bounded search query and publishes only source-backed seller records.

Features

• One to 100 unique seller offers per run.
• Up to ten input URLs, ten discovered products, five search pages, and three concurrent browser tasks.
• Product and seller ratings, current/original prices, discount and savings, availability, delivery, COD, return policy, EMI, categorized offers, public product details, and provenance.
• Optional exact brand/seller, price, rating, discount, Flipkart Assured, and stock filters.
• Direct, Apify, or custom proxy transport with bounded pacing, timeouts, and retries.
• Stable record IDs, compact rows with no null or empty top-level fields, and product/seller deduplication.
• Fail-closed challenge, login, geofence, rate-limit, WAF, device, timeout, and unrecognized-source handling.

The Actor does not log in, solve CAPTCHAs, reuse protected sessions, or collect private data.

Input modes

Product URLs

{
  "mode": "productUrl",
  "productUrls": [
    "https://www.flipkart.com/example-product/p/itm0000000000000?pid=MOBEXAMPLE1234"
  ],
  "maxItems": 25
}

Direct sellers pages

{
  "mode": "sellersPageUrl",
  "sellersPageUrls": [
    "https://www.flipkart.com/sellers?pid=MOBEXAMPLE1234"
  ],
  "includeProductDetails": false
}

Search

{
  "mode": "search",
  "query": "smartphone",
  "maxPages": 1,
  "maxProducts": 3,
  "maxItems": 25
}

Use the Input tab for all bounds, filters, detail/media/rating toggles, pacing, timeout, retry, and proxy options. pincode is optional reference context only and is not submitted to Flipkart.

Dataset and run output

Each dataset item is a sellerOffer backed by the public Flipkart page. Core identity and provenance fields are always present; optional fields are omitted when Flipkart does not disclose them. Failure diagnostics are never mixed into the dataset.

The OUTPUT key-value record reports atomic completion status, source/filter counts, task and navigation counts, proxy mode, bounded failure evidence, and runtime. A filter that matches no source rows is a successful valid-empty run. A blocked, ambiguous, or extraction-failed run fails closed with an empty dataset.

Local verification

npm ci --ignore-scripts
npm run lint
npm test
apify validate-schema
apify run --input-file test-apify-input.json --purge

Flipkart’s public layout can change. When it does, the Actor intentionally fails closed instead of publishing guessed seller data.