Google Keyword Suggest — Long-Tail Keywords & Autocomplete

Scrape Google autocomplete for any seed keyword — A–Z expansion, question modifiers, prefix expansion. Hundreds of long-tail keywords per seed, with country and language targeting. No API key.

What you get

• One row per autocomplete suggestion, deduplicated across all expansion queries
• Relevance score for every suggestion (Google's internal ranking, 0–1300+) — sort by this for the most popular long-tails
• Country and language targeting — get suggestions as a US English searcher, a UK searcher, a Brazilian Portuguese searcher, etc.
• Five expansion modes — pick the volume / cost trade-off that fits your task
• Direct export to JSON, CSV, Excel, or Google Sheets

Use cases

• SEO content briefs — feed a seed topic and get a question-keyword cluster ("how does X work", "what is X", …) for FAQ sections
• Long-tail keyword research — surface 100s of low-competition phrases per topic without paying $99/mo for Ahrefs or AnswerThePublic
• PPC keyword discovery — find buyer-intent modifiers (buy X, X near me, X reviews) by running suffix + prefix expansion
• Product/feature ideation — pull all autocompletes for your product category to see what users actually ask
• AI/LLM seed data — generate diverse query lists for evals, RAG benchmarks, or prompt-injection tests
• Trend tracking — schedule a daily run and diff today's suggestions vs. yesterday's to spot emerging searches

Expansion modes

Mode	What it fires	Queries per seed
none	Just the seed	1
suffix-a-z	seed, seed a, seed b, … seed z	27
prefix-a-z	seed, a seed, b seed, … z seed	27
both-a-z	Suffix + prefix combined	53
questions	Seed plus who/what/when/where/why/how/can/are/is/do/does/will/should/which	15
all	Suffix + prefix + questions — maximum coverage	67

Each query returns up to 15 suggestions; deduplication usually leaves 100–300 unique long-tails per seed in all mode.

How to use

1. Enter one or more Seed keywords (e.g. crm software, personal trainer, kubernetes)
2. Pick an Expansion mode — start with suffix-a-z (cheap, broad coverage)
3. Set Country and Language for geo/locale-targeted autocomplete (e.g. us + en, br + pt, de + de)
4. Set Max items (0 = no cap) and run
5. Sort the dataset by relevance descending — top hits are what real users search most

Output format

{
  "seed": "crm software",
  "query": "crm software for",
  "suggestion": "crm software for small businesses",
  "relevance": 560,
  "type": "QUERY",
  "expansionType": "suffix-a-z",
  "expansionToken": "f",
  "country": "us",
  "language": "en",
  "fetchedAt": "2026-05-18T01:30:00.000Z"
}

The query field is the exact string sent to Google (seed + the expansion token); suggestion is what Google returned. relevance is Google's internal ranking — higher = more searched.

Input options

Field	Type	Description
Seeds	array	One or more seed keywords
Expansion mode	enum	none / suffix-a-z / prefix-a-z / both-a-z / questions / all
Country	string	Two-letter country code (us, uk, de, br, in, …)
Language	string	Two-letter language code (en, es, pt, fr, de, …)
Max items	integer	Cap on suggestions across all seeds. 0 = no cap

Notes

• Suggestions are deduplicated per seed (case-insensitive)
• type is Google's classification of the suggestion — QUERY for normal searches, NAVIGATION for site jumps, ENTITY for known entities
• Country and language combinations should be plausible (gl=br hl=pt works; gl=br hl=ja will fall back to defaults)