Wikipedia Search Suggest Scraper

Get live Wikipedia search suggestions for any seed keyword — the exact phrases Wikipedia shows people as they type. The fastest way to see real demand in Wikipedia's own words.

Why Wikipedia suggestions?

Suggestions are ranked by actual user behaviour, not by a third-party estimate. They tell you:

• 🔍 What people actually type — not what a keyword tool guesses
• 🏷️ Long-tail phrases you would never brainstorm on your own
• 📈 Demand ranking — position 1 is the strongest signal
• ⚡ Fresh data — pulled live at run time, not from a stale index

Typical users: researchers, content teams and knowledge-graph builders.

How it works

For every seed keyword the Actor queries Wikipedia's public suggestion endpoint. With a–z expansion enabled (default) it also queries seed a … seed z, uncovering far more long-tail results than a single lookup. Results are deduplicated and pushed to the dataset.

Input

Field	Type	Default	Description
keywords	array	—	Seed keywords, one per line
language	select	en	Language edition to query (en, de, fr, es, tr, ja, ru …).
expandAlphabet	boolean	true	Also query "seed a" … "seed z"
expandNumbers	boolean	false	Also query "seed 0" … "seed 9"
keepDuplicates	boolean	false	Keep a suggestion found under several prefixes
requestDelayMs	integer	120	Politeness delay between requests

Output

One dataset item per suggestion:

{
    "seed": "machine learning",
    "query": "machine learning r",
    "suggestion": "machine learning reel wall mount",
    "position": 2,
    "language": "en",
    "collectedAt": "2026-08-05T09:00:00.000Z"
}

position is the rank in Wikipedia's own list — 1 means the strongest demand signal for that query.

What to expect

One seed with a–z expansion ≈ 27 queries. A batch of 10 seeds usually returns a few hundred to a few thousand unique suggestions in under two minutes.

FAQ

Is this legal? The Actor collects publicly available, non-personal suggestion data — the same suggestions any visitor sees on Wikipedia. No login, no personal data, no access controls bypassed.

Do suggestions include search volume? Wikipedia does not expose volume. Use position as the demand-rank signal, or combine with your favourite volume tool.

Can I run it on a schedule? Yes — use Apify Schedules to track how suggestions change over time.

Can I export to CSV / Google Sheets? Yes, from the run's Output tab or via the API.