An Amazon search API that answers in ~100 milliseconds.
Every other Amazon search API scrapes a results page while you wait — 2 to 50 seconds per query, priced per attempt. Ours answers from a continuously refreshed warehouse of 3.5M+ products: keyword in, ranked product shortlists out, grouped by category or price tier, with brand facets and demand signals. No Amazon approval, no Associates account, no live-scrape latency. Try it below — the demo hits the production endpoint.
not a scrape queued behind a proxy
refreshed continuously — 10 more marketplaces shortly
— deterministic, no LLM in the path
Search it yourself. Watch the response time.
Every search below hits our production endpoint as you click — nothing pre-rendered. Broad queries (try toothbrush) come back grouped by product category; specific ones (try graphics card) come back as Budget / Mid-range / Premium price tiers. The grouping is chosen algorithmically.
Live responses from our production warehouse — observed snapshots, not Amazon's live feed. Demo is limited to 30 searches per hour per visitor and two products per group; the API itself returns more, with full facets and filters. Request a test key for unrestricted access.
Coverage note: the search index currently covers Amazon's best-seller (BSR) product universe and excludes media products (books, music, video). We're working on extending it to include media and products beyond the best-seller range.
See a raw JSON response (real output, trimmed)
GET /v1/shopping?q=toothbrush&country=us
x-api-key: your_test_key
{
"query": "toothbrush", "country": "us", "currency": "USD",
"total": 1571, "mode": "category",
"groups": [
{ "category": "Health & Household > Oral Care > ... > Powered Toothbrushes > Ultrasonic",
"count": 110,
"products": [
{ "asin": "B072YVWBXH",
"title": "Aquasonic Black Series Ultra Whitening Toothbrush – ADA Accepted...",
"brand": "Aquasonic", "price": 35.95, "rating": 4.6,
"review_count": 136657, "in_stock": true, "is_fba": true,
"image_url": "https://m.media-amazon.com/images/I/71sZ7M1OHwL._AC_SY355_.jpg",
"amazon_url": "https://www.amazon.com/dp/B072YVWBXH" } ] },
{ "category": "Health & Household > Oral Care > ... > Manual Toothbrushes",
"count": 98,
"products": [
{ "asin": "B00CC6XSSQ",
"title": "Colgate Extra Clean Toothbrush, Soft Toothbrush for Adults, 6 Count",
"brand": "Colgate", "price": 4.96, "rating": 4.6,
"review_count": 68604, "in_stock": true, "is_fba": false, "..." : "..." } ] }
],
"facets": {
"brands": [
{ "value": "Oral-B", "count": 102, "avg_rating": 4.56, "total_ratings": 641540, "est_monthly_sold": 14109 },
{ "value": "Colgate", "count": 44, "avg_rating": 4.64, "total_ratings": 261848, "est_monthly_sold": 3337 }
]
}
}Why it responds in milliseconds when everyone else takes seconds.
Every conventional Amazon search API works the same way: your query goes into a queue, a proxy fetches an Amazon results page, a parser extracts what it can, and you get an answer 2 to 50 seconds later — billed whether the parse succeeded or not. That architecture cannot be fast, because the scrape happens after you ask.
We inverted it. Our crawlers continuously walk Amazon's best-seller universe and maintain a ranked, deduplicated search index of the 3.5M+ products people actually buy — refreshed around the clock, variant families collapsed to one representative product, accessories filtered out of head queries. When your query arrives, the work is already done. The index answers in ~100ms with results ranked by rating, review mass, and availability, grouped the way a merchandiser would group them.
The trade-off, stated plainly: responses are observed snapshots with timestamps, not Amazon's live feed at the moment of your request. Best-sellers refresh most often; the long tail less so. Media products (books, music, video) are currently excluded — coverage of media and of products beyond the best-seller range is in the works. For purchase-moment accuracy on a specific ASIN, an on-demand fresh fetch is available. For search, discovery, ranking, and comparison — the things a search API is actually for — the warehouse is the right architecture, and the speed is the proof.
What a search response includes.
- Ranked, grouped shortlists Broad queries return the best products per category (field-collapsed, best-per-type). Specific queries return Budget / Mid-range / Premium price tiers with real percentile boundaries. Accessory noise is filtered automatically — a search for "graphics card" returns GPUs, not HDMI cables.
- Full product records ASIN, title, brand, price, rating, review count, in-stock status, FBA flag, image URL, full category path, and a direct Amazon link — per product, in one response.
- Brand and category facets with demand signal Top brands and categories in the result set, each with product count, average rating, total ratings, and estimated monthly units sold where Amazon exposes it. Enough to build a category page, not just a list.
- US, UK, and Walmart today amazon.com, amazon.co.uk, and walmart.com in one API with one schema. Germany, Canada, Japan, France, Italy, Spain, India, Australia, Mexico, and Brazil are next — shortly available.
- Per-ASIN depth behind every result Any product in a search response can be expanded via the product data API: multi-seller offers with seller identity, and optional price / BSR / seller-count history reaching back up to ~2 years.
vs. the alternatives.
| WDES | Live-scrape APIs | Amazon Creators API | |
|---|---|---|---|
| Typical search response | ~100 ms | 2–50 seconds | Fast, when it works |
| Eligibility | API key — that's it | API key | 10 qualifying sales / 30 days, ongoing |
| Result shape | Ranked, grouped, accessory-filtered | Raw SERP rows | ~10 raw results |
| Failed requests | Not billed | Often billed / retried on your credits | Free, but 500s reported under load |
| Brand / demand facets | Included | No | No |
| Multi-seller offers & seller identity | Yes | Separate endpoint, per-page credits | Not available |
| Price / BSR / seller history | Up to ~2 years, same API | Not held | Not available |
If you qualify for the Creators API and stay qualified, use it — it's free and it's Amazon's live catalog. We exist for everyone who can't, and for the capabilities it never had. Switching from a specific provider? See vs. Rainforest, vs. Keepa, vs. Canopy.
Who builds on it.
- AI assistants, agents & MCP tools A deterministic product-search tool call that returns in ~100ms, structured for LLM consumption. Grouped shortlists an agent can reason over, not 40 raw SERP rows to re-rank. Agents make repeated tool calls; sub-second latency is the difference between usable and not.
- Affiliate & content sites "Best X for Y" pages, comparison tables, and product boxes that fill themselves from a keyword — with ratings, review counts, and price tiers ready to render. No Associates eligibility required to read the data. Lost PA-API? Start here.
- Price-comparison, deal & shopping apps Interactive search that keeps up with a user typing. In-stock filtering, price tiers, and Walmart cross-marketplace results in the same response schema.
- Market research & category intelligence Brand facets with estimated monthly sold, category distributions, and the per-ASIN history behind any result — a category snapshot in one call instead of a thousand scrapes.
Get a free test key.
Tell us what you're building — an agent tool, a content site, a comparison app — and we'll set you up with a test key and the endpoint docs the same day.
Request a test key