Official Python SDK for Amazon Scraper API. Pay only for successful (2xx) responses — non-2xx never cost you anything. Pricing starts at $0.90 per 1,000 requests on pay-as-you-go, drops to a flat $0.50 per 1,000 requests on Custom plans. 1,000 free requests on signup, no card. Drop into any Python project to fetch structured Amazon product data, run keyword searches, or queue async batches with webhook callbacks.
Measured on our own infrastructure against a 30-query mixed international set:
| Metric | Value |
|---|---|
| Median latency (product, US) | ~2.6 s |
| P95 latency | ~6 s |
| P99 latency | ~10.5 s |
| Price / 1,000 Amazon products | $0.50 flat |
| Concurrent threads (entry paid plan) | 50 |
| Marketplaces supported | 20+ |
| Billing unit | per successful (2xx) response |
pip install amazonscraperapi-sdkRequires Python >= 3.9. Built on httpx.
from amazonscraperapi import AmazonScraperAPI
asa = AmazonScraperAPI(api_key="asa_live_...")
product = asa.product(query="B09HN3Q81F", domain="com")
print(product["title"])
# "LEGEND COOKWARE 5-Ply Stainless Steel Cookware Set, 14-Piece..."
print(product["price"], product["currency"])
# 549.99 USD
print(product["rating"], product["reviews_count"])
# 4.5 865{
"asin": "B09HN3Q81F",
"title": "LEGEND COOKWARE 5-Ply Stainless Steel Cookware Set, 14-Piece...",
"brand": "LEGEND COOKWARE",
"price": 549.99,
"price_strikethrough": None,
"currency": "USD",
"rating": 4.5,
"reviews_count": 865,
"stock": "In Stock",
"is_prime": True,
"images": ["https://m.media-amazon.com/images/I/71Q3a8JH5NL._AC_SY355_.jpg", "..."],
"bullet_points": "5-PLY STAINLESS STEEL COOKWARE HEATS EVENLY - Five layers of steel...",
"sales_rank": [
{"ladder": [{"name": "Kitchen & Dining", "url": None}], "rank": 36177},
{"ladder": [{"name": "Kitchen Cookware Sets", "url": None}], "rank": 174},
],
}results = asa.search(
query="wireless headphones",
domain="co.uk",
sort_by="avg_customer_review",
pages=1,
)
for p in results["products"]:
# Sponsored cards have sponsored_position instead of organic_position.
print(p["organic_position"], p["asin"], p["title"], p["price"], p["rating"])batch = asa.create_batch(
endpoint="amazon.product",
items=[
{"query": "B09HN3Q81F", "domain": "com"},
{"query": "B000ALVUM6", "domain": "de", "language": "de_DE"},
# ... up to 1,000
],
webhook_url="https://your.server/webhooks/asa",
)
print("batch id:", batch["id"])
# SAVE THIS. Webhook signing secret is returned only once:
print("webhook secret:", batch["webhook_signature_secret"])
# Alternative: poll
status = asa.get_batch(batch["id"])
print(f"{status['processed_count']}/{status['total_count']}")from amazonscraperapi import verify_webhook_signature
from fastapi import FastAPI, Request, HTTPException
import json, os
app = FastAPI()
@app.post("/webhooks/asa")
async def asa_webhook(request: Request):
raw = await request.body()
signature = request.headers.get("X-ASA-Signature")
if not verify_webhook_signature(signature, raw, secret=os.environ["WEBHOOK_SECRET"]):
raise HTTPException(401, "invalid signature")
payload = json.loads(raw)
# process payload["results"]Building a production-grade Amazon scraper in-house is a 2-4 week engineering project plus permanent maintenance. This SDK wraps Amazon Scraper API, which has already solved:
| Pain point | What we handle |
|---|---|
| Amazon CAPTCHAs / robot pages | Auto-detected, retried through a heavier proxy tier (datacenter, residential, premium) |
| Brittle CSS selectors | Extractors update as Amazon changes layouts. Your code stays the same. |
| 20+ marketplaces | amazon.de, .co.uk, .co.jp, .com.br, and more. Each with marketplace-specific parsing quirks. |
| Country-matched residential IPs | amazon.de auto-routes through German IPs. |
| Rotating proxies + anti-fingerprinting | TLS fingerprints, headers, cookies handled server-side. |
| Structured JSON output | Title, price, rating, reviews, variants, seller, images. Parsed. |
| Batch/async jobs | 1,000 ASINs submitted, webhook-delivered on completion. |
Time saved: a greenfield Python Amazon scraper built to this spec takes roughly 80 engineer-hours (including anti-bot handling and marketplace variants). This SDK is 10 minutes.
Every failed call raises AmazonScraperAPIError with status_code, code (the API's error code, or None) and body (the parsed JSON body). The SDK does not retry on its own, so match on e.code and decide:
import time
from amazonscraperapi import AmazonScraperAPI, AmazonScraperAPIError
asa = AmazonScraperAPI(api_key="asa_live_...")
try:
product = asa.product(query="B09HN3Q81F", domain="com")
except AmazonScraperAPIError as e:
if e.code == "INSUFFICIENT_CREDITS":
pass # top up
elif e.code == "RATE_LIMITED":
time.sleep(e.body["error"].get("retry_after_ms", 1000) / 1000)
elif e.code == "target_unreachable":
time.sleep(60) # not charged; retry once
else:
raise| HTTP | code |
When you see it | Recommended action |
|---|---|---|---|
| 400 | invalid_params |
Missing query, unsupported domain, bad sort_by |
Fix request; don't retry |
| 401 | INVALID_API_KEY |
Missing, malformed, revoked | Verify ASA_API_KEY; rotate if leaked |
| 402 | INSUFFICIENT_CREDITS |
Balance empty | Top up; renews each billing cycle |
| 404 | product_not_found |
ASIN delisted, malformed, or not on this marketplace | Don't retry; drop the ASIN |
| 429 | RATE_LIMITED |
Over request-rate budget | Wait error.retry_after_ms; retry |
| 429 | CONCURRENCY_LIMIT |
Over plan's parallel cap | Reduce parallelism or upgrade. X-Concurrency-* headers guide backoff |
| 502 | target_unreachable |
Amazon blocked every internal retry (last_block_reason says why) |
Retry after 60s or more. Not charged |
| 502 | extraction_failed |
Layout we can't parse | Retry later; report with X-Request-Id if it persists. Not charged |
| 503 | SERVICE_OVERLOADED |
Global circuit breaker | Honor Retry-After: 60. Rare |
| 500 | INTERNAL_ERROR |
Our bug | Report with X-Request-Id |
The default request timeout is 120s (the API's own ceiling); override it with AmazonScraperAPI(api_key=..., timeout=...).
Flat-credit promise: non-2xx responses are free. A basic request costs 5 credits (billing unit; end-customer price is $0.90 per 1,000 basic requests PAYG). Future JS-rendered calls will cost 15 credits. Every response has X-Request-Id for traceability — quote it in any support ticket.
app.amazonscraperapi.com. 1,000 free requests on signup, no credit card required.
- Website: https://www.amazonscraperapi.com/
- Docs: https://amazonscraperapi.com/docs
- Pricing: https://amazonscraperapi.com/pricing
- Node SDK: amazon-scraper-api-sdk · Go SDK: github.com/ChocoData-com/amazon-scraper-api-sdk-go · CLI: amazon-scraper-api-cli · MCP server: amazon-scraper-api-mcp
MIT