Autocomplete is the cheapest keyword data there is. Every suggestion is something real people typed often enough for the engine to offer it. The annoying part is collecting it: seven engines, seven response formats, and the alphabet trick (seed + a, seed + b...) means dozens of requests per keyword.
I built a small Apify Actor that does the boring part. You send seeds, it returns every suggestion from Google, YouTube, Bing, DuckDuckGo, Amazon, eBay and the App Store, deduplicated and tagged with search intent.
curl -X POST "https://api.apify.com/v2/acts/vaquilla~keyword-autocomplete-explorer/run-sync-get-dataset-items?token=YOUR_APIFY_TOKEN" \
-H "Content-Type: application/json" \
-d '{"seeds":["running shoes"],"engines":["google","youtube","amazon"],"expansions":["alphabet","questions"]}'
Each row looks like this:
{ "keyword": "best running shoes for flat feet", "engine": "google", "position": 3, "intent": "commercial", "isQuestion": false }
Three things I use it for
-
Content briefs. Turn on
questionsand you get the "how", "what" and "why" searches around a topic. Those become H2s. - Amazon listings. The Amazon engine returns what buyers type, which is rarely what sellers write in titles.
- Feeding an AI agent. Give the agent real demand data before it writes, so it stops inventing keywords.
Numbers and limits
In my test run it returned 1,677 keywords for one seed in about 4 seconds. It costs $0.40 per 1,000 keywords and there is no subscription. It does not return search volume, only the suggestions and their position.
Link: Google, YouTube & Amazon Autocomplete Scraper on Apify
If an engine or a country you need is missing, tell me and I will add it.
Top comments (0)