Amazon Seller Intelligence

Turn a handful of product keywords into a list of third-party Amazon sellers with their legal business name, registered address, email, phone and fulfillment model (FBA/FBM).

Why this exists

Amazon publishes a surprising amount of business data on every third-party seller: the registered company name, the full legal address, often an email address and a phone number. The problem is that this data lives on thousands of individual seller pages that nobody wants to open by hand.

If you sell software or services to e-commerce merchants, that data is a lead list. If you run brand protection, it tells you who is actually behind a storefront. If you do market research, it tells you where the sellers in a niche are based and whether they ship themselves or rely on Amazon's warehouses.

This actor does the tedious part. You give it keywords, it returns clean, deduplicated seller records.

Who uses it

Who	What they do with it
B2B sales teams (logistics, shipping, SaaS)	Build lists of merchant-fulfilled sellers with a direct email or phone
Brand owners and IP lawyers	Identify the legal entity behind a storefront selling their products
Market researchers	Map how many sellers compete in a niche, by country and fulfillment model
Aggregators and investors	Source FBA businesses in a category, with size signals from feedback counts
Agencies	Find sellers with their own web shop (from business emails, live domain checks and web search)

Quick start

1. Enter a few product keywords, the way a shopper would type them.
2. Pick the marketplace (amazon.com, .co.uk, .de, .fr, .it, .es or .nl).
3. Set the maximum number of sellers you want.
4. Run it. Results land in the dataset, one row per seller.

That's the whole setup. The defaults (deep discovery, fulfillment detection, China-based sellers excluded) are the ones most users want.

Example output

{
  "sellerId": "A1B2C3D4E5F6G7",
  "marketplace": "amazon.de",
  "sellerName": "AlpineGear Shop",
  "businessName": "AlpineGear GmbH",
  "businessAddress": "Industriestrasse 12 | 82467 Garmisch-Partenkirchen | DE",
  "countryCode": "DE",
  "email": "service@alpinegear.de",
  "phone": "+49 8821 123456",
  "websiteDomain": "alpinegear.de",
  "websiteSource": "email_domain",
  "sellerUrl": "https://www.amazon.de/sp?seller=A1B2C3D4E5F6G7",
  "storefrontUrl": "https://www.amazon.de/s?me=A1B2C3D4E5F6G7",
  "about": "Family-run outdoor equipment retailer since 1998.",
  "fulfillmentModel": "fbm_only",
  "fbaRatio": 0.0,
  "offersChecked": 4,
  "feedbackCountLifetime": 12480,
  "feedbackCount12Months": 1893,
  "positivePercent12Months": 97.0,
  "productAsins": ["B08XYZ1234", "B09ABC5678"],
  "keywordsMatched": ["ski helmet"],
  "productsSeen": 3,
  "hasContact": true,
  "error": false
}

Output fields

Field	Description
sellerId	Amazon merchant ID, unique per marketplace
sellerName	Public storefront display name
businessName	Registered legal company name
businessAddress	Registered business address
countryCode	Two-letter country code parsed from the address
email, phone	Contact details published on the seller profile, when present
websiteDomain	The seller's own web shop domain, when one could be found
websiteSource	How the web shop was found: email_domain, brand_probe or web_search
sellerUrl	Link to the seller's public profile page
fulfillmentModel	fba_only, fbm_only, mixed, unknown or not_checked
fbaRatio	Share of sampled offers fulfilled by Amazon (0–1)
feedbackCountLifetime, feedbackCount12Months, positivePercent12Months	Seller size and reputation signals
productAsins	Sample of products the seller was seen offering
keywordsMatched	Which of your keywords surfaced this seller
hasContact	True when at least one direct contact channel was found

Input options worth knowing

• deepDiscovery (default on) fetches every seller's full profile page after finding them. This is what gets you addresses, emails and phone numbers. Turn it off for a faster, cheaper run that returns seller identity and fulfillment model only — note that without profiles there is no registered address, so the country filter and China exclusion cannot apply.
• checkFulfillment (default on) classifies each seller as FBA, FBM or mixed from the offers seen during discovery, at no extra cost. Useful when you specifically target merchant-fulfilled sellers.
• discoverWebsite (default on) looks for each seller's own e-commerce site: business email domain first, then a direct check of brand-name domains, then a web search. Sellers with their own shop are often the most valuable rows in the list.
• excludeChineseSellers (default on) drops sellers whose registered address points to mainland China, Hong Kong, Taiwan or Macau.
• sellerCountryFilter keeps only sellers registered in one country, e.g. DE.
• maxSellers caps the number of paid results. The run also respects the spending limit you set on the platform, whichever is lower.

Reliability

• Sellers are deduplicated by merchant ID across all your keywords.
• Failed lookups never land in your dataset by default, so you don't pay for errors. They are stored in the FAILED_ITEMS record of the run's key-value store, together with a RUN_SUMMARY record with coverage statistics.
• Transient fetch errors are retried automatically before being reported.
• If the whole run cannot fetch anything, it fails with a clear message and you are not charged for results.

Limitations

• Email and phone coverage depends on what each seller publishes. Expect an email or phone on a meaningful share of sellers, not on all of them. hasContact tells you instantly which rows have one.
• Country parsing relies on the registered address format; a small share of sellers come back with countryCode: null. These rows pass any country filter, so review them manually if the filter matters to you.
• With strict filters (country filter plus China exclusion in a mixed niche) the run can return fewer sellers than maxSellers. Discovery over-fetches to compensate, but it cannot invent sellers that do not exist in the niche.
• Fulfillment classification is based on the offers observed during discovery. It is a strong signal, not a guarantee; sellers whose observed offers carry no shipping origin come back unknown.
• Website discovery is heuristic. websiteSource tells you how each domain was found: email_domain and brand_probe are strong signals, web_search is a good lead worth a quick check before outreach.
• Supported marketplaces are amazon.com, .co.uk, .de, .fr, .it, .es and .nl.
• All data comes from publicly visible Amazon pages. Nothing is scraped from behind a login.