How to Do Amazon Keyword Research with Python

Amazon keyword research starts with one question: what do shoppers actually type into the search bar? The most direct answer is Amazon's own autocomplete. When you type "coffee b", Amazon suggests "coffee beans", "coffee bean grinder", and "coffee bar accessories". Those suggestions are coming from real search behavior.

Most free Amazon keyword research tools are built on the same data, but they limit how many lookups you can run, and it's hard for you to integrate them into your own workflow. In this tutorial, you'll use the Amazon Autocomplete API from SerpApi to get these suggestions as a clean JSON or Markdown output, expand one seed keyword into dozens of long-tail terms, and save them to a CSV file.

What data does the Amazon Autocomplete API return?

Each request returns a suggestions array. Each suggestion includes:

value: the suggested search term, for example coffee beans.
type: the suggestion type. KEYWORD is a regular search suggestion. WIDGET is a grouped block, such as a price-range shortcut.
items: for WIDGET suggestions, a list of options, each with its own value and amazon_link.
thumbnail: a small product image for the suggestion, when Amazon shows one.
amazon_link: the Amazon search results URL for that suggestion.
serpapi_link: a ready-made SerpApi request that gets autocomplete suggestions for this suggestion. Useful for going one level deeper.
serpapi_amazon_link: a ready-made Amazon Search API request for this suggestion, so you can see which products rank for it.

Here's the result from the trimmed response for coffee in JSON

{
  "suggestions": [
    {
      "value": "Coffee table by price",
      "type": "WIDGET",
      "items": [
        {
          "value": "Under $100",
          "amazon_link": "https://www.amazon.com/s?k=coffee+table&amp=&rh=n%3A1063318%2Cp_n_price%3A8570764011&amp=&rnid=8570763011&amp=&nav_sdd=aps&amp=&crid=26CJG6W215CUX&amp=&sprefix=coffee&amp=&ref=nb_sb_ss_w_sbl-tr-t1_k0_1_6_0"
        },
        ...
        ...
        ...
      ]
    },
    {
      "value": "coffee table",
      "type": "KEYWORD",
      "thumbnail": "https://m.media-amazon.com/images/I/41Q0IEeqGBL.__AC_SL75__.jpg",
      "serpapi_link": "https://serpapi.com/search.json?amazon_domain=amazon.com&engine=amazon_autocomplete&k=coffee+table",
      "serpapi_amazon_link": "https://serpapi.com/search.json?amazon_domain=amazon.com&device=desktop&engine=amazon&k=coffee+table",
      "amazon_link": "https://www.amazon.com/s?k=coffee%20table"
    },
    ...
    ...
    ...
 }

SerpApi returns Amazon's search suggestions in a structured output

Getting started with SerpApi

You need a free SerpApi account before you can use the API. You can upgrade to a paid plan later if you need more searches, faster speeds, or additional features.

First, create an account and verify your email. Then grab your API key from your account dashboard. The free plan includes 250 searches per month, and you can try any query live in the interactive playground before writing code.

You should store your API key in a safe location if you're sharing or publishing your code. If it's ever leaked, you can generate a new one from the dashboard. The examples below read the key from an environment variable.

Install the SerpApi library (optional)

SerpApi has official libraries for Python, JavaScript, Ruby, Java, and more. They're a thin wrapper around the API and aren't required. The API works just as well with plain GET requests, cURL, or fetch() in Node.js.

For the Python examples, install the official client:

pip install serpapi

Review the Amazon Autocomplete API documentation

The engine is amazon_autocomplete, and the text you'd type into Amazon's search bar goes in the k parameter. Other useful parameters:

amazon_department: limit suggestions to one department. Defaults to aps (All Departments). See the full list of Amazon departments.
amazon_domain: the Amazon marketplace to use, such as amazon.co.uk or amazon.de. Defaults to amazon.com. See the supported Amazon domains.
output: set to md for a token-efficient Markdown response.

For the full field reference and live examples, see the Amazon Autocomplete API documentation.

SerpApi's Amazon Autocomplete API documentation

How to scrape Amazon autocomplete suggestions

The results are identical across every library, so use whichever method fits your stack.

GET request

This gets suggestions for "coffee":

https://serpapi.com/search.json?engine=amazon_autocomplete&k=coffee&api_key=SERPAPI_API_KEY

By default you get JSON, but you can request Markdown by adding output=md (or using the /search.md endpoint, or the Accept: text/markdown header). Markdown returns the data in a more token-efficient format built with tables and links, which is handy when feeding results to an LLM or AI agent:

https://serpapi.com/search.json?engine=amazon_autocomplete&k=coffee&output=md&api_key=SERPAPI_API_KEY

Python

This gets suggestions for "coffee" and prints each keyword with its type, using the official SerpApi Python library:

import os
import serpapi

client = serpapi.Client(api_key=os.environ["SERPAPI_API_KEY"])

results = client.search({
    "engine": "amazon_autocomplete",
    "k": "coffee",
})

for suggestion in results.get("suggestions", []):
    print(f"{suggestion['value']} | {suggestion['type']}")

Output:

Coffee table by price | WIDGET
coffee table | KEYWORD
coffee pods | KEYWORD
coffee | KEYWORD
coffee maker | KEYWORD
coffee creamer | KEYWORD
coffee beans | KEYWORD
coffee grinder | KEYWORD
coffee filters | KEYWORD
coffee bean grinder | KEYWORD
coffee machine | KEYWORD

JavaScript and Node.js

This runs the same request with the SerpApi JavaScript library:

import { getJson } from "serpapi";

const results = await getJson({
  engine: "amazon_autocomplete",
  api_key: process.env.SERPAPI_API_KEY,
  k: "coffee",
});

for (const suggestion of results.suggestions ?? []) {
  console.log(`${suggestion.value} | ${suggestion.type}`);
}

cURL

This gets suggestions straight from the command line:

curl --get https://serpapi.com/search \
 -d api_key="SERPAPI_API_KEY" \
 -d engine="amazon_autocomplete" \
 -d k="coffee"

Filter suggestions by Amazon department and marketplace

The suggestions for a broad word mix with unrelated products. For example, "coffee" in All Departments, Amazon suggests "coffee table" and "coffee grinder" alongside "coffee pods".

You can set amazon_department to narrow the results. With amazon_department=grocery, the same query returns only food and drink terms: "coffee pods", "coffee creamer", "coffee syrup", "coffee k cups", "coffee mate", and "coffee substitutes".

https://serpapi.com/search.json?engine=amazon_autocomplete&k=coffee&amazon_department=grocery&api_key=YOUR_API_KEY

To research another marketplace, use amazon_domain with any of the supported Amazon domains. Suggestions reflect what shoppers search for on that site:

https://serpapi.com/search.json?engine=amazon_autocomplete&k=coffee&amazon_domain=amazon.co.uk&api_key=YOUR_API_KEY

Build a free Amazon keyword research tool with Python

Each request returns about 10 suggestions. To build a real keyword list, use the same trick as most Amazon keyword suggestion tools by adding each letter of the alphabet after your seed keyword. For example, "coffee a", "coffee b", "coffee c", and so on; each returns a different set of long-tail suggestions.

Here's what the script below does:

Runs the seed keyword plus one query per letter (27 queries in total)
Keeps only KEYWORD suggestions and skips WIDGET blocks
Removes duplicates
Saves the list to amazon_keywords.csv, with the query that found each keyword

import csv
import os
import string

import serpapi

client = serpapi.Client(api_key=os.environ["SERPAPI_API_KEY"])

SEED = "coffee"
DEPARTMENT = "grocery"   # use "aps" for All Departments
MAX_QUERIES = 27         # 1 seed + 26 letters; each query uses one search


def get_suggestions(query):
    results = client.search({
        "engine": "amazon_autocomplete",
        "k": query,
        "amazon_department": DEPARTMENT,
    })
    return [
        s["value"]
        for s in results.get("suggestions", [])
        if s.get("type") == "KEYWORD"
    ]


queries = [SEED] + [f"{SEED} {letter}" for letter in string.ascii_lowercase]
queries = queries[:MAX_QUERIES]

keywords = {}
for query in queries:
    for keyword in get_suggestions(query):
        keywords.setdefault(keyword.lower(), query)

with open("amazon_keywords.csv", "w", encoding="UTF-8", newline="") as f:
    writer = csv.writer(f)
    writer.writerow(["keyword", "source_query"])
    for keyword, source in sorted(keywords.items()):
        writer.writerow([keyword, source])

print(f"Ran {len(queries)} queries and saved {len(keywords)} unique keywords to amazon_keywords.csv")

💡

MAX_QUERIES caps the number of searches per run. Lower it while testing so you don't use up your monthly searches.

Amazon keywords research output in CSV format

There are a few ways you can extend this tools:

Add modifiers. Besides letters, try words shoppers use to narrow a search, such as "for", "best", "organic", or "with".
Go one level deeper. Run the script again with a top suggestion as the new seed, for example "coffee beans".
Compare marketplaces. Add amazon_domain and run the same seed on amazon.com and amazon.co.uk.

Check which products rank for each keyword

A keyword list tells you what shoppers search for. For Amazon product keyword research, the next step is to see what they find. Every KEYWORD suggestion includes a serpapi_amazon_link, which is a ready-made Amazon Search API request for that term. You can call it to get the ranking products, their prices, ratings, and review counts.

For a full walkthrough of the Amazon Search API, pagination, and product details, see our guide on how to scrape Amazon product data below:

FAQ

Does Amazon have a keyword tool?

Amazon doesn't offer a public keyword research tool for shoppers. Sellers can see some search term data in Seller Central, but the most accessible source of real search terms is the autocomplete dropdown in Amazon's search bar, which is what this API returns.

What are keywords in Amazon?

Amazon keywords are the words and phrases shoppers type into Amazon's search bar. Sellers use them in product titles, bullet points, backend search terms, and ad campaigns so their products show up for those searches.

Is there a free Amazon keyword research tool?

Several free tools exist, but most limit how many lookups you can run. The script in this post is a free alternative for up to 250 searches per month on SerpApi's free plan. Each query uses one search.

Does the Amazon Autocomplete API show search volume?

No. It returns the suggestions Amazon shows, in the order Amazon shows them, but not search volume. Use the suggestions to find which terms exist, then check volume with a separate keyword tool if you need it.

Conclusion

The Amazon Autocomplete API turns Amazon's search suggestions into structured JSON and Markdown output, so you can run Amazon keyword research in bulk instead of typing into the search bar one query at a time. You can then pair it with the Amazon Search API to see which products rank for each keyword, or the Amazon Product API to track prices on the products you find.

If you need help getting started, contact us. We're happy to help.

How to Do Amazon Keyword Research with Python

Noraina Nordin

What data does the Amazon Autocomplete API return?

Getting started with SerpApi

Install the SerpApi library (optional)

Review the Amazon Autocomplete API documentation

How to scrape Amazon autocomplete suggestions

GET request

Python

JavaScript and Node.js

cURL

Other languages and no-code solutions

Filter suggestions by Amazon department and marketplace

Build a free Amazon keyword research tool with Python

Check which products rank for each keyword

FAQ

Does Amazon have a keyword tool?

What are keywords in Amazon?

Is there a free Amazon keyword research tool?

Does the Amazon Autocomplete API show search volume?

Conclusion

Free Plan · 250 searches / month

APIs

Easy Integrations

Features

Use Cases

Company

Pricing

What data does the Amazon Autocomplete API return?

Getting started with SerpApi

Install the SerpApi library (optional)

Review the Amazon Autocomplete API documentation

How to scrape Amazon autocomplete suggestions

GET request

Python

JavaScript and Node.js

cURL

Other languages and no-code solutions

Filter suggestions by Amazon department and marketplace

Build a free Amazon keyword research tool with Python

Check which products rank for each keyword

FAQ

Does Amazon have a keyword tool?

What are keywords in Amazon?

Is there a free Amazon keyword research tool?

Does the Amazon Autocomplete API show search volume?

Conclusion

Free Plan · 250 searches / month