Keyword Suggestions from Google, YouTube, Bing and Baidu Autocomplete

What it does

Autocomplete is one of the most direct signals of what people search for: every suggestion is a query real users type. Collecting it by hand means typing a seed, then the seed plus each letter of the alphabet, in every search engine you care about. The Keyword Suggestions actor does that in bulk.

Give it seed keywords and pick engines: Google, YouTube, Bing and Baidu. With a-z expansion (the default) each seed is also sent as "seed a" to "seed z", which usually returns a few hundred unique long-tail keywords per seed and engine. questions expansion adds who, what, why, how, can, is and similar words in front of the seed; prepositions adds for, with, vs, near and others after it. Results are deduplicated across the expansion.

You get one row per seed and engine with each suggestion's text, position and the exact query that produced it, or a flat row per keyword for spreadsheets. It costs $0.30 per 1,000 seeds per engine without expansion and $1 per 1,000 with expansion, billed per seed and engine rather than per keyword. Empty and rate-limited rows are free.

How it works

1. You provide seeds, choose engines and an expansion mode, and set country and language for the market.
2. Each seed is expanded into its query set (1 request with none, 27 with a-z, 14 with questions, 13 with prepositions) and sent to each engine's public suggest endpoint.
3. Requests are paced politely: at most 2 in flight and at least 350 ms apart per host, with back-off on HTTP 429.
4. Suggestions are deduplicated per seed and engine (case- and whitespace-insensitive) and capped at maxSuggestionsPerSeed.
5. Each complete seed and engine row is charged once; rows with no suggestions, partial failures or rate limits are free.

Input example

A typical run input (JSON). Every field is also available as a form in the Apify console.

{
  "seeds": [
    "coffee maker",
    "running shoes"
  ],
  "engines": [
    "google",
    "youtube",
    "bing"
  ],
  "expansion": "a-z",
  "country": "us",
  "language": "en",
  "dedupe": true,
  "outputMode": "grouped"
}

Output example

In the sample each row is one seed on one engine, with the suggestion count and the keywords found by the a-z expansion.

Pricing

Pay per event: you are charged only for the results you receive, with no subscription. Set a maximum total charge on any run and it stops cleanly at the cap.

Prices are read from the live Apify Store listing when this page is built. The listing is always authoritative.

Use cases

How it compares

Autocomplete tells you what people type, not how often. If you need search volumes, difficulty or CPC, a paid keyword suite is the right tool; this actor is a cheap way to generate the candidate list to feed into one.

Limits and honest notes

• Sources: Google and YouTube suggestions come from Google's public suggest endpoint (suggestqueries.google.com, which has no robots.txt), Bing from its OpenSearch suggestion feed (api.bing.com/osjson.aspx, not disallowed by robots), Baidu from suggestion.baidu.com. Re-checked 2026-10-20.
• Not included, on purpose: Amazon, DuckDuckGo and Naver (their robots.txt disallows it), eBay (robots plus an explicit ban on automated access), Yahoo (its terms ban automated access without permission) and Yandex (its search terms ban automated requests outside its paid API).
• Polite by design: an honest user agent (ActorForge-KeywordSuggestions/1.0), at most 2 requests in flight and at least 350 ms between requests per host, no proxies and no proxy rotation. So an a-z seed takes about 10 seconds per engine; Google and YouTube share one host, Bing and Baidu run in parallel with it.
• Rate limits: on HTTP 429 (or Google's "sorry" page) the actor waits 5 s, then 20 s. If the host keeps answering 429, or answers 401/403 even once, that engine is stopped for the rest of the run and its remaining rows come back as status rate-limited, free. Google can rate-limit shared cloud IPs; just re-run later.
• Billing is per seed x engine, never per keyword: flat mode (one row per suggestion) costs exactly the same as grouped mode. Only complete rows with at least one suggestion (status ok) are charged; no-suggestions, partial (some queries failed), rate-limited and error rows are free.
• Suggestions are what each engine shows for that query, country and language at run time. They are not search volumes. Google partly localises by the requester's IP as well as gl, so a few results can reflect the data-centre location.
• Each response has up to about 10 suggestions (Bing up to ~20, Baidu ~10, often fewer for long-tail or non-Chinese seeds). Dedupe (on by default) keeps the first occurrence across the expansion; maxSuggestionsPerSeed caps a row.
• The run's spend limit is respected: each seed x engine is reserved before its requests start and charged before its row is saved, and the run stops cleanly at the cap.

FAQ

Use it from code

Run the actor through the Apify API with the input above.

cURL

curl -X POST "https://api.apify.com/v2/acts/actor-forge~keyword-suggestions/run-sync-get-dataset-items" \
  -H "Authorization: Bearer $APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"seeds": ["coffee maker", "running shoes"], "engines": ["google", "youtube", "bing"], "expansion": "a-z", "country": "us", "language": "en", "dedupe": true, "outputMode": "grouped"}'

JavaScript (Node.js)

// Node.js 18+ (built-in fetch). Set APIFY_TOKEN in your environment.
const input = {
  "seeds": [
    "coffee maker",
    "running shoes"
  ],
  "engines": [
    "google",
    "youtube",
    "bing"
  ],
  "expansion": "a-z",
  "country": "us",
  "language": "en",
  "dedupe": true,
  "outputMode": "grouped"
};

const res = await fetch(
  "https://api.apify.com/v2/acts/actor-forge~keyword-suggestions/run-sync-get-dataset-items",
  {
    method: "POST",
    headers: {
      Authorization: `Bearer ${process.env.APIFY_TOKEN}`,
      "Content-Type": "application/json",
    },
    body: JSON.stringify(input),
  }
);
const items = await res.json();
console.log(items.length, items[0]);

Python

# pip install apify-client   (set APIFY_TOKEN in your environment)
import os
from apify_client import ApifyClient

client = ApifyClient(os.environ["APIFY_TOKEN"])
run_input = {
    "seeds": [
        "coffee maker",
        "running shoes"
    ],
    "engines": [
        "google",
        "youtube",
        "bing"
    ],
    "expansion": "a-z",
    "country": "us",
    "language": "en",
    "dedupe": True,
    "outputMode": "grouped"
}

run = client.actor("actor-forge/keyword-suggestions").call(run_input=run_input)
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item)

The run-sync-get-dataset-items endpoint waits for the run (up to 5 minutes) and returns the dataset rows directly. For longer runs, start the run asynchronously and read the dataset when it finishes, as the Python example does. Replace APIFY_TOKEN with your own token from the Apify console.

Related actors

More SEO & Marketing Intelligence tools · All actors