Amazon Deals API: Lightning Deals, Savings and End Times as JSON

To get Amazon deals as JSON, call the SellerMagnet /api/amazon-deals endpoint with a marketplace ID and, optionally, a department. One request returns up to 50 current deals with the deal price, the list price it is measured against, the savings percentage, start and end times and a lightning-deal flag. Poll it every few minutes, keep a set of the dealIds you have seen, and you have a deal feed that costs one credit per poll.

What does the deals API return?

The endpoint reads the marketplace's deals page and returns its deal collection as a list. Prices are in the marketplace currency (dealPrice.priceCurrency), times are UTC, and savings_percentage is taken from the deal badge or, when the badge shows none, computed from the two prices. coupon_savings carries the badge text when there is one, such as 20 % Rabatt on amazon.de.

One deal from /api/amazon-deals on amazon.de (example values)

{
  "success": true,
  "data": {
    "deals": [
      {
        "dealId": "5675cf93",
        "asin": "B094YCTK6P",
        "productTitle": "IceUnicorn Krabbelschuhe Baby Lauflernschuhe",
        "dealLink": "https://www.amazon.de/dp/B094YCTK6P",
        "dealPrice": {"priceTotal": 15.89, "priceCurrency": "EUR"},
        "basisPrice": 18.69,
        "savings_percentage": 15,
        "coupon_savings": null,
        "dealStartTime": "2025-07-08T04:15:00.000Z",
        "dealEndTime": "2025-07-08T16:15:00.000Z",
        "isLightningDeal": true,
        "image_url": "https://images-eu.ssl-images-amazon.com/images/I/717dw8J-hbL.jpg"
      }
    ]
  }
}

How do I monitor Amazon deals with Python?

Poll the endpoint, compare with the dealIds stored from the last poll, and report two things: deals that are new, and deals that end within the next two hours. The state lives in a small JSON file, so a cron job can run the script every 15 minutes without a database.

deal_watch.py - new deals and deals ending soon, one credit per run

import datetime as dt
import json
import os
import pathlib

import requests

API = "https://sellermagnet-api.com/api/amazon-deals"
STATE = pathlib.Path("deals_seen.json")
ENDING_SOON = dt.timedelta(hours=2)


def fetch_deals(marketplace_id: str, category_id: str | None = None, count: int = 50) -> list:
    params = {"marketplaceId": marketplace_id, "count": count,
              "api_key": os.environ["SELLERMAGNET_API_KEY"]}
    if category_id:
        params["category_id"] = category_id
    resp = requests.get(API, params=params, timeout=90)
    try:
        body = resp.json()
    except ValueError:  # e.g. an HTML error page from a proxy
        body = {}
    if resp.status_code != 200 or not body.get("success"):
        raise RuntimeError(f"{resp.status_code} {body.get('message')}")
    return body["data"]["deals"]


def ends_in(deal: dict) -> dt.timedelta | None:
    end = deal.get("dealEndTime")
    if not end or end == "N/A":
        return None
    return dt.datetime.fromisoformat(end.replace("Z", "+00:00")) - dt.datetime.now(dt.timezone.utc)


seen = set(json.loads(STATE.read_text())) if STATE.exists() else set()
deals = fetch_deals("A1PA6795UKMFR9", category_id="electronics")

for deal in deals:
    price = deal["dealPrice"]["priceTotal"]
    basis = deal.get("basisPrice")            # number, or null when the page shows no list price
    key = deal["dealId"] if deal.get("dealId") not in (None, "N/A") else deal["asin"]
    left = ends_in(deal)
    if price <= 0:
        continue                              # price could not be read on this poll
    if key not in seen:
        print(f"NEW  {deal['asin']} {price} {deal['dealPrice']['priceCurrency']}"
              f" (was {basis}, -{deal.get('savings_percentage')}%) {deal['productTitle'][:50]}")
    if left is not None and dt.timedelta(0) < left <= ENDING_SOON:
        print(f"ENDS {deal['asin']} in {int(left.total_seconds() // 60)} min")

STATE.write_text(json.dumps(sorted({d["dealId"] if d.get("dealId") not in (None, "N/A") else d["asin"]
                                    for d in deals})))

How do I filter deals by department?

Pass category_id with a department slug - the documented example is electronics; the top-level IDs in the category explorer are the candidates for other departments. Without it you get the marketplace's whole deal collection - lightning deals and the deal of the day alike. An unknown category_id is not rejected; it is passed to Amazon's deals filter as it is, so a typo simply returns fewer or no deals. count defaults to 30 and stops at 50; a higher value returns HTTP 400 Count Max: 50 after the credit is charged. geo_location (a postcode) is best-effort, as on search.

Deals ending within two hours, from the shell

curl -sG "https://sellermagnet-api.com/api/amazon-deals" \
  --data-urlencode "marketplaceId=A1PA6795UKMFR9" \
  --data-urlencode "category_id=electronics" \
  --data-urlencode "count=50" \
  --data-urlencode "api_key=$SELLERMAGNET_API_KEY" \
  | jq --arg now "$(date -u +%Y-%m-%dT%H:%M:%S)" \
       --arg limit "$(date -u -v+2H +%Y-%m-%dT%H:%M:%S 2>/dev/null || date -u -d '+2 hours' +%Y-%m-%dT%H:%M:%S)" \
       -r '.data.deals[] | select(.dealEndTime != "N/A" and .dealEndTime > $now and .dealEndTime <= $limit)
           | [.asin, .dealPrice.priceTotal, .savings_percentage, .dealEndTime] | @tsv'

What does an empty deals list mean?

An empty deals array with success: true means the deals page carried a deal collection with nothing in it for that filter - a niche department at a quiet hour - and it is billed like any other answer. A page that could not be read at all is not turned into an empty list: it returns HTTP 502 or 503, which the error-handling guide covers. Poll every 10 to 15 minutes: deal windows are hours long, and a lightning deal that appears and sells out inside one interval is not worth a credit a minute.

To judge a deal, pair it with the product's recorded prices: the price history endpoint shows whether the deal price really is a low, and the bestsellers endpoint tells you how the product sells in its category.

Bottom line: one credit per poll, deduplicate by dealId, parse the times as UTC and skip entries whose price could not be read. The deals endpoint page shows the full response, and a free account includes 500 credits - five days of 15-minute polling.