Keyword Suggestions Scraper

What does the Keyword Suggestions Scraper do?

The Keyword Suggestions Scraper collects public autocomplete text for seed keywords from Google, Bing, Yahoo, Amazon, Etsy, YouTube, Pinterest, Wikipedia, and Answers.com. It uses each source's own public suggestion route, with a bounded browser fallback for Etsy and Pinterest. It does not estimate search volume or keyword difficulty.

Why use the Keyword Suggestions Scraper?

Use autocomplete phrases for topic discovery, product naming, and content planning. Each successful keyword/platform pair produces one dataset record containing the ordered suggestions, a one-based rank for each suggestion, and the actual extraction host and route. The Actor preserves a stable source order, removes duplicates, and limits the list independently for each pair. Its OUTPUT key-value record distinguishes successful, empty, blocked, unavailable, timed-out, malformed, and failed source requests.

The Actor can run through Apify scheduling and API integrations. It can use an authorized Apify Proxy configuration or custom proxy URLs when the user enables one; proxies are off by default and do not bypass source access controls.

What data can it extract?

Group	Fields	Purpose
Identity	id, recordType, keyword, normalizedKeyword, keywordLength, platform, platformName	Stable record key and seed/platform identity.
Suggestions	suggestions, suggestionItems, suggestionCount, firstSuggestion	Ordered text and per-item position, prefix match, and completion. Pinterest's browser fallback is labeled as a related-search refinement.
Locale and provenance	language, countryCode, source, sourcePageUrl, extractionUrl, endpointUrl	Requested locale and the host/URLs associated with the record. Locale support varies by source.
Request details	extractionMethod, suggestionType, status, httpStatus, apiRequestAttempts, retryCount, maxItemsRequested, responseContentType, responseBytes, requestDurationMs, requestedAt, scrapedAt	How the source was reached, bounded response metadata, and UTC timing. responseBytes is omitted for browser fallback records.

Only successful, non-empty source responses become dataset records. Empty and failed source outcomes remain in the OUTPUT key-value summary rather than being represented as fabricated suggestion rows.

How to scrape keyword suggestions

1. Enter one or more seed keywords in the Actor's Input tab.
2. Choose platforms, locale, and the per-source suggestion limit.
3. Start the Actor and inspect the dataset for suggestion records. Use OUTPUT to review sources that were empty or unavailable.
4. Download the dataset or use its API URL for downstream analysis.

How much will it cost to scrape keyword suggestions?

The number of source operations is approximately the number of unique keywords multiplied by the number of selected platforms. Retries add requests after temporary failures; Etsy and Pinterest may also use browser work after their API route fails. Increasing concurrency can reduce elapsed time while increasing simultaneous requests. The maxItems cap limits stored suggestions per pair. Check the run's compute usage in Apify Console; this Actor does not promise a fixed price.

Input

This input uses the current schema and passes the Actor's runtime validator:

{
  "keywords": ["software", "ai tools"],
  "platforms": ["google", "bing", "etsy", "pinterest", "wiki"],
  "language": "en",
  "countryCode": "US",
  "maxItems": 20,
  "maxConcurrency": 3,
  "maxRequestRetries": 2,
  "requestTimeoutSecs": 30,
  "useApifyProxy": false
}

Section	Field	Type and default	Behavior
Search	keywords	String list; required; 1–100 values	Unique, non-empty seeds; each is trimmed and whitespace is collapsed.
Search	platforms	String list; all nine sources by default	Allowed values: google, bing, yahoo, amazon, etsy, youtube, pinterest, wiki, answers.
Locale	language	String; en	Two- or three-letter language code, optionally followed by an uppercase country code.
Locale	countryCode	String; US	Two-letter country code. Sources may ignore it.
Result limits	maxItems	Integer; 20; range 1–50	Maximum suggestions for each keyword/platform pair.
Developer options	maxConcurrency	Integer; 3; range 1–9	Maximum source pairs handled at once.
Developer options	maxRequestRetries	Integer; 2; range 0–5	Retries temporary and other retryable API failures. Permanent access errors are not retried.
Developer options	requestTimeoutSecs	Integer; 30; range 5–120	Timeout for each API or browser operation.
Developer options	useApifyProxy	Boolean; false	Enables an authorized Apify Proxy session.
Developer options	proxyGroups	String list; empty	Optional groups used only when Apify Proxy is enabled.
Developer options	proxyCountryCode	String; unset	Optional two-letter Apify Proxy country.
Developer options	proxyUrls	String list; empty	Authorized HTTP(S) proxy URLs; these take precedence over Apify Proxy.

Unknown input fields are rejected. There is no public debug or fixture input option.

The complete input schema, including accepted platform values and limits, is in .actor/input_schema.json.

Output

The dataset contains 29 declared fields, including optional API response bytes. The suggestionItems entries retain source order and contain position, text, startsWithKeyword, and completion. A completion is derived only when the returned suggestion begins with the normalized seed; otherwise it is null.

The dataset and run-summary schemas are .actor/dataset_schema.json, .actor/key_value_store_schema.json, and .actor/output_schema.json.

The OUTPUT key-value record contains the overall status, timing and counts, proxy-use flag, and one bounded source summary per requested keyword/platform pair. A source block is not reported as an empty success. source is the host that supplied the record, so sources that share infrastructure (such as Google and YouTube suggestions) are represented accurately.

You can download the dataset in various formats such as JSON, HTML, CSV, or Excel.

Source routes

Each platform is read from its own current public suggestion route, with no third-party host in between:

Platform	Route	Payload field used
Google	suggestqueries.google.com/complete/search?client=firefox	[1][]
Bing	api.bing.com/osjson.aspx	[1][]
Yahoo	search.yahoo.com/sugg/gossip/gossip-us-ura/	[1][]
Amazon	completion.amazon.com/api/2017/suggestions	suggestions[].value
Etsy	www.etsy.com/suggestions_ajax.php?search_query=…&search_type=all&pathname=/search	results[].query
YouTube	suggestqueries.google.com/complete/search?client=youtube	[1][][0]
Pinterest	www.pinterest.com/resource/AdvancedTypeaheadResource/get/	resource_response.data.items[].query where type is query
Wikipedia	<lang>.wikipedia.org/w/api.php?action=opensearch	[1][]
Answers.com	www.answers.com/question/search?q=	data[].title

Etsy's response also carries one shop-name suggestion whose query is HTML markup, and Pinterest's item list blends query completions with user creator profiles; neither is a keyword suggestion, so both are excluded rather than stored.

Current Chrome spot-check

On 2026-09-23, I entered software followed by a space in the visible search box in Chrome. These are examples directly visible in current UI suggestion lists; this was a page inspection, not an Actor run:

Public page	Visible suggestions
Google search	software download, software examples, software company madurai, software testing, software development
Bing search	software center, software update, software download, software fix, software engineering
Yahoo search	software download, software engineering, software adalah, software wikipedia, software update
Amazon search	software defined radio, software for pc, software engineering, software library, software downloads
YouTube search	software video, software engineering, software developer movie in hindi, software developer short film, software update

That inspection did not show a text-completion dropdown for Etsy, Pinterest, Wikipedia, or Answers.com on their search pages, which is why the Actor reads those sources' own suggestion routes instead of their rendered pages. Each source's status is still reported in OUTPUT, and an empty or unavailable source is never filled with text from another platform.

Tips and advanced options

• Keep keyword batches small while comparing platforms because each keyword is requested from each selected source.
• Use maxConcurrency, finite retries, and a reasonable timeout to keep request volume bounded.
• Suggestions vary with time, language, country, source, and account/session state; they are not search-volume data.
• The Actor does not solve CAPTCHAs, bypass sign-in prompts, or turn unrelated product cards into autocomplete suggestions. Etsy's DOM fallback accepts only listbox options that start with the seed. Pinterest's page fallback is explicitly marked as a refinement source. The Etsy and Pinterest browser fallback runs only after their API route fails; a validated empty API answer is reported as an empty source.
• You can review run history, logs, the dataset API URL, and the OUTPUT key-value record in the Apify Console. Use the Actor's Issues tab for bugs or feature requests and include a run ID plus a short description.

FAQ, support, and responsible use

Why is a platform missing? A source can return no suggestions, be unavailable, or restrict automated requests. Check its status in OUTPUT; an unavailable result is not replaced with a phrase from another platform.

Does the Actor report search volume? No. Autocomplete text is a discovery signal and does not include search-volume estimates, CPC, or competition metrics.

Use only public information and follow each site's terms, robots guidance, rate limits, and applicable law. Do not use this Actor to collect private personal data. Results can still contain personal data; process them only with an appropriate legal basis. This Actor is independent and is not affiliated with Google, Microsoft/Bing, Yahoo, Amazon, Etsy, YouTube, Pinterest, Wikipedia, or Answers.com.