🔑 Semrush Keyword Magic Tool – Keyword Research, Volume Estimates & Search Intent

The Semrush Keyword Magic Tool actor expands a single seed keyword into a full list of related keyword variations and returns an estimated SEO metric set for each one: monthly search volume, CPC range, competition index, search intent, SERP feature type and a set of derived opportunity scores. Give it a keyword, a country code and a language code, and it returns a structured keyword research dataset you can sort, filter and plan content against.

Two things are worth stating plainly before you use this keyword research tool. The keyword list itself is expanded from Google's public autocomplete suggestions for your seed term, country and language — so the variations are real queries people type. The accompanying metrics are estimates generated deterministically by the actor's own model from the keyword text, country and language, not live figures pulled from a paid keyword API. They are consistent and repeatable for the same input, and they are useful for relative comparison and prioritisation across a keyword set, but they should not be treated as verified search volumes. Where an exact volume number matters commercially, confirm it in Google Keyword Planner or your own analytics.

---

📊 What Data Can You Extract with This Keyword Magic Tool?

Each dataset item is one keyword variation. The fields group into six categories:

Category	Fields	What it tells you
🔤 Keyword	keyword	The keyword variation, expanded from autocomplete suggestions for your seed term
📈 Volume	avgMonthlySearches, monthlySearchVolumes	Estimated average monthly search volume, plus a twelve-month array of month, year and estimated searches
💰 Commercial value	lowCPC, highCPC, monetizationScore	Estimated cost-per-click range as a formatted currency string, and a derived monetisation score combining CPC and competition
⚔️ Competition	competitionIndex, competitionValue	A numeric competition index and its banded label — low, medium or high
🎯 Intent & SERP	intent, intentConfidence, serpFeatureType, estimatedCTR	Classified search intent with a confidence value, the SERP feature expected for the query, and an estimated click-through rate
🧭 Opportunity signals	contentGapScore, keywordFreshness, advice	Derived content-gap and freshness scores, plus a plain-English advice string summarising what the metrics imply

The most immediately actionable field is advice. Rather than making you interpret six numeric scores, it converts them into a short recommendation — for example Keyword freshness is low: produce updated content regularly to maintain relevance. For a content team triaging a hundred-keyword list, that column is usually the fastest route from data to a decision.

---

🌟 Key Features of the Keyword Magic Tool

Feature	Description
🌱 Seed keyword expansion	One seed term expands into a list of related keyword variations drawn from Google's public autocomplete suggestions
🌍 Country targeting	The country field takes a two-letter code — us, uk, in, ca, au, de, fr and others — and localises both suggestions and estimates
🗣️ Language targeting	The language field takes a two-letter code such as en, es, fr, de, pt, it or ja
📊 Twelve-month volume array	monthlySearchVolumes returns a month-by-month series, making seasonality visible rather than hiding it behind an annual average
🎯 Intent classification	Keywords are classified as informational, commercial or transactional, with an intentConfidence value attached
🔍 SERP feature prediction	serpFeatureType indicates the feature likely to dominate the result page — Featured Snippet, Image Pack, Video Pack, People Also Ask or None
🧮 Volumes snapped to Keyword Planner steps	Estimated volumes are snapped to the standard Google Ads Keyword Planner buckets (10, 20, 30, 50, 70, 90, 110 … 1,000,000) so they read like planner output
♻️ Deterministic and repeatable	The same keyword, country and language always produce the same metrics, so runs are reproducible and diffable
📤 Configurable result count	maxResults caps how many keyword variations are returned, defaulting to 100

---

🚀 Why Choose This Keyword Magic Tool?

Real autocomplete-driven keyword discovery. The variation list comes from Google's public suggestion endpoint for your seed, country and language. Those are queries people genuinely type, which makes them a sound basis for topic and content planning regardless of how you treat the metric columns.

Deterministic, comparable metrics. Because every metric is derived consistently from the keyword, country and language, two runs of the same input produce identical output. That makes the dataset useful for relative prioritisation — ranking keywords against each other within one research set — and easy to version-control alongside a content plan.

Seasonality exposed, not averaged away. monthlySearchVolumes gives you twelve month-year-searches entries per keyword. A keyword whose estimated volume peaks in spring is a very different editorial commitment from one that is flat all year, and an annual average hides that entirely.

A single, simple input. Four fields, three of them with sensible defaults. There is no proxy configuration, no account to connect and no API key to manage — you type a seed keyword and run.

---

📥 Input

The Semrush Keyword Magic Tool actor requires a seed keyword, a country and a language.

{
  "keyword": "free ai image generator",
  "country": "us",
  "language": "en",
  "maxResults": 100
}

🔧 Keyword Magic Tool Input Fields

Field	Type	Required	Default	Description
keyword	string	✅ Yes	—	The main keyword or phrase to research. Examples: free ai image generator, digital marketing, best laptops 2024
country	string	✅ Yes	"us"	Two-letter country code for localised keyword data. Examples: us, uk, in, ca, au, de, fr
language	string	✅ Yes	"en"	Two-letter language code. Examples: en, es, fr, de, pt, it, ja
maxResults	integer	❌ No	100	Maximum number of keyword variations to extract

💡 Input Examples

UK English content research — moderate list for an editorial calendar:

{
  "keyword": "project management software",
  "country": "uk",
  "language": "en",
  "maxResults": 150
}

German-market keyword expansion:

{
  "keyword": "elektroauto laden",
  "country": "de",
  "language": "de",
  "maxResults": 100
}

Quick exploratory run using defaults:

{
  "keyword": "digital marketing",
  "country": "us",
  "language": "en"
}

---

📤 Output

One dataset item per keyword variation. This is a real record from an actual run of the keyword magic tool:

{
  "keyword": "free ai image generator",
  "avgMonthlySearches": 6600,
  "lowCPC": "$0.93",
  "highCPC": "$4.55",
  "competitionIndex": 41,
  "competitionValue": "medium",
  "intent": "informational",
  "intentConfidence": 0.45,
  "advice": "Keyword freshness is low: produce updated content regularly to maintain relevance.",
  "contentGapScore": 0.544,
  "estimatedCTR": 0.08,
  "keywordFreshness": 0.2,
  "serpFeatureType": "Featured Snippet",
  "monetizationScore": 4.49,
  "monthlySearchVolumes": [
    { "month": "July", "year": 2025, "searches": 6600 },
    { "month": "August", "year": 2025, "searches": 6600 }
  ]
}

🧾 Keyword Magic Tool Output Fields — Keyword Metrics

Field	Type	Description
keyword	string | null	The keyword variation this record describes
avgMonthlySearches	integer | null	Estimated average monthly search volume, snapped to Keyword Planner steps
lowCPC	string | null	Estimated low end of the cost-per-click range, formatted as currency
highCPC	string | null	Estimated high end of the cost-per-click range, formatted as currency
competitionIndex	integer | null	Numeric competition index for the keyword
competitionValue	string | null	Banded competition label: low, medium or high
intent	string | null	Classified search intent — informational, commercial, transactional, or a combination
intentConfidence	number | null	Confidence value attached to the intent classification
advice	string | null	Plain-English recommendation derived from the keyword's metrics
contentGapScore	number | null	Derived content-gap score for the keyword
estimatedCTR	number | null	Estimated organic click-through rate
keywordFreshness	number | null	Derived freshness score indicating how time-sensitive the query is
serpFeatureType	string | null	SERP feature expected for the query, e.g. Featured Snippet, Image Pack, Video Pack, People Also Ask, None
monetizationScore	number | null	Derived score combining CPC and competition into a single monetisation signal
monthlySearchVolumes	array | null	Twelve entries of {month, year, searches} showing estimated seasonality

🧾 Run Configuration Fields

These fields reflect the configuration the run executed with:

Field	Type	Description
country	string | null	Country code the run targeted
language	string | null	Language code the run targeted
maxResults	integer | null	Maximum result count configured for the run
outputFile	string | null	Output file name used by the run

---

💻 How to Use the Keyword Magic Tool (Step by Step)

Step 1: Choose a Seed Keyword with Real Expansion Potential

The seed keyword determines the entire result set, so choose a head term with genuine variation beneath it rather than an already-long phrase. project management software will expand into dozens of useful modifiers; best free project management software for small remote teams is already so specific that autocomplete has little left to add. If you are unsure, type the term into Google and look at how many suggestions appear — that is roughly what the expansion has to work with.

Step 2: Set the Country and Language Codes to Match Your Audience

country and language are separate fields for a reason: you might target Spanish-language content in the United States, or English content in Germany. Both take two-letter codes. The pair affects which autocomplete suggestions are returned and how volume and CPC estimates are localised, so getting them right materially changes your output.

Step 3: Set maxResults for the Depth You Need

The default of 100 keyword variations is a good working size for a content plan. Raise it when you are building a comprehensive topic cluster and want the long tail; lower it when you are quickly sanity-checking whether a topic has any depth at all. The value is a ceiling — if autocomplete produces fewer suggestions than you asked for, the run returns what it found.

Step 4: Run the Keyword Magic Tool

Press Start. The actor first collects autocomplete suggestions for your seed across country and language, then computes the metric set for each variation. The log reports how many candidate keywords were generated and how many were processed. Runs are quick because there is no browser and no proxy layer involved.

Step 5: Sort and Triage the Keyword Dataset

Open the Dataset tab and sort. A sensible first pass is to sort descending by avgMonthlySearches, then scan competitionValue for low and medium entries — those are the volume-versus-difficulty sweet spots. Then read the advice column on your shortlist, since it condenses the remaining scores into a recommendation.

Step 6: Cluster Keywords by Intent Before Writing

Group your shortlist on intent. Informational keywords belong in guides, explainers and comparison content; commercial keywords belong on category and review pages; transactional keywords belong on product and pricing pages. Mixing intents on one page is one of the most common reasons a well-researched piece of content fails to rank, and intentConfidence tells you how firmly each classification is held.

Step 7: Cross-Check Volumes Before Committing Budget

Because the metric columns are model-generated estimates rather than live third-party API figures, use them to rank and prioritise, not to forecast revenue. Before committing paid budget or a major content investment to a specific term, verify its volume and CPC in Google Keyword Planner or your own Search Console data. Use this tool for breadth and prioritisation; use your own data for the final number.

---

🔌 API Access & Integrations

Run the keyword magic tool from your own code. The synchronous endpoint starts a run and returns the dataset in one call:

curl -X POST "https://api.apify.com/v2/acts/scrapers-hub~semrush-keyword-magic-tool/run-sync-get-dataset-items?token=YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "keyword": "free ai image generator",
    "country": "us",
    "language": "en",
    "maxResults": 100
  }'

With the official Python client:

from apify_client import ApifyClient

client = ApifyClient("YOUR_TOKEN")

run_input = {
    "keyword": "project management software",
    "country": "uk",
    "language": "en",
    "maxResults": 150,
}

run = client.actor("scrapers-hub/semrush-keyword-magic-tool").call(run_input=run_input)

for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["keyword"], item["avgMonthlySearches"], item["competitionValue"], item["intent"])

The actor works with Apify's standard integrations for Zapier, Make, Google Sheets, Slack and generic webhooks, so a finished keyword research run can populate a content calendar sheet automatically.

---

💡 Best Use Cases for Keyword Research Data

📝 Content Calendar Planning

Expand a pillar topic into a hundred variations, then sort by avgMonthlySearches and filter to competitionValue of low or medium. The resulting shortlist, grouped by intent, is effectively a quarter's editorial calendar with each article's angle already decided by its intent classification.

🗺️ Topic Cluster Construction

Use the seed keyword as your pillar page and the returned variations as cluster articles. Grouping variations by shared intent and serpFeatureType shows you which subtopics belong together and which need their own dedicated page rather than a section within an existing one.

📅 Seasonal Content Timing

monthlySearchVolumes returns a twelve-entry series per keyword. Identify which months show peaks for your priority terms and publish six to eight weeks ahead of them, since content typically needs lead time to build rankings before demand arrives.

💵 PPC Keyword Shortlisting

lowCPC, highCPC, competitionIndex and monetizationScore together give a first-pass view of which terms look commercially attractive. Treat this as a shortlisting exercise: narrow a large keyword universe down to a manageable candidate set here, then validate the survivors in your ad platform before bidding.

🎯 Search Intent Mapping

Filter the dataset by intent and map each group to a page type. This exercise routinely reveals that a site has ten informational articles competing with each other and no transactional page at all for the terms that would actually convert.

🔍 SERP Feature Targeting

serpFeatureType tells you what dominates the result page. Terms flagged as Featured Snippet reward concise, directly-answering paragraphs near the top of the page; Image Pack terms reward strong original visuals; Video Pack terms may need video content rather than another written article.

🌐 International Keyword Expansion

Run the same seed with different country and language pairs to see how a topic's vocabulary changes across markets. Direct translation of keywords is a reliable way to miss local phrasing, and comparing autocomplete-derived variation lists per market avoids that trap.

---

⚙️ Tips for Better Keyword Research Results

• Seed broad, filter narrow. Head terms expand into far more useful variation sets than long-tail seeds. Start wide and use maxResults plus post-run filtering to get to your shortlist.
• Match country and language to your actual audience, not your own locale. These fields change both the suggestion set and the estimates, and defaulting to us/en out of habit is a common source of irrelevant results.
• Treat the metrics as relative, not absolute. They are model-generated and deterministic. Ranking keywords against each other within one dataset is sound; quoting an exact volume figure to a client is not.
• Run several related seeds. Three or four seeds around the same theme produce far better topical coverage than one seed with a very high maxResults, because autocomplete depth per seed is finite.
• Read advice before the numeric columns. It condenses freshness, volume and CPC into a single recommendation, which is much faster to triage across a long list.
• Keep runs alongside your content plan. Because output is deterministic for a given input, storing the dataset with the plan means anyone can reproduce exactly the research the plan was built on.

---

🛠️ Troubleshooting

I got fewer keywords than maxResults. maxResults is a ceiling. Autocomplete only has so many suggestions for a given seed, country and language, and narrow or unusual seeds naturally produce short lists. Use a broader head term if you need more depth.

The volumes do not match Google Keyword Planner. They will not match exactly. The keyword variations are drawn from Google autocomplete, but the metrics are estimates generated by the actor's model rather than live planner figures. Use them for relative prioritisation and verify specific numbers in Keyword Planner before spending against them.

Running the same keyword twice gives identical results. That is intentional. Metrics are deterministic for a given keyword, country and language, which makes runs reproducible and comparable over time.

All my results are in English despite setting a different country. country and language are separate fields. Setting country to de without also setting language to de will target the German market in English. Set both.

advice is empty for some keywords. The advice string is assembled from whichever conditions apply to that keyword. When none of the triggering conditions are met, the field comes back empty — which itself indicates a keyword with no particular flags against it.

---

❓ Frequently Asked Questions About Keyword Research

What does the Semrush Keyword Magic Tool actor do? It expands one seed keyword into a list of related keyword variations using Google's public autocomplete suggestions for your chosen country and language, and returns an estimated SEO metric set for each variation.

Where does the keyword list come from? From Google's public autocomplete suggestion endpoint, queried for your seed keyword, country code and language code. These are real suggested queries.

Where do the search volume and CPC figures come from? They are estimates generated deterministically by the actor from the keyword text, country and language — not live data from a paid keyword API. They are consistent and repeatable, and best used for comparing keywords against each other rather than as verified absolute figures.

Is this connected to a Semrush account? No. There is no account to connect and no API key to supply. The input is a keyword, a country code and a language code.

How many keyword variations can I get in one run? As many as maxResults allows and autocomplete can supply. The default is 100.

Which countries and languages are supported? Both fields take free-text two-letter codes. Common values are documented in the field descriptions — us, uk, in, ca, au, de, fr for country and en, es, fr, de, pt, it, ja for language.

Why are search volumes round numbers like 480, 720 and 6600? Volumes are snapped to the standard Google Ads Keyword Planner steps, which is the same bucketing the planner itself uses, so the figures read naturally alongside planner exports.

What is monthlySearchVolumes for? It returns twelve {month, year, searches} entries per keyword so you can see estimated seasonality rather than only an annual average — useful for timing publication ahead of demand peaks.

How is search intent classified? Keywords are classified as informational, commercial or transactional based on the modifiers they contain, with intentConfidence recording how firmly the classification is held. A keyword can carry more than one intent label.

What does contentGapScore mean? It is a derived opportunity score returned per keyword. Like the other derived scores, it is most useful for ranking keywords within a single dataset rather than as a standalone number.

Can I run several seed keywords at once? The input takes one keyword per run. To research several seeds, start several runs — easily automated through the API or an Apify schedule.

Do I need a proxy for this keyword magic tool? No. The actor makes direct requests with no proxy layer and no browser.

What export formats are available? Apify datasets export as JSON, CSV, Excel, XML, RSS and HTML, and are also readable through the Apify API and official clients.

Should I set PPC budgets from the CPC figures? No. Use lowCPC and highCPC to shortlist candidate terms, then validate the survivors in your ad platform before committing spend.

Can I schedule recurring keyword research runs? Yes, through Apify's scheduler. Bear in mind that output for a given input is deterministic, so scheduled runs are most useful when you vary the seed keyword rather than repeating one.

---

🆘 Support & Feedback

If the keyword magic tool behaves unexpectedly — an empty result set for a seed you know has suggestions, or an obviously wrong locale response — open a ticket on the Issues tab of this actor and include the input JSON and the run ID.

Need a custom variant, such as different metric derivations, additional locales, or bulk seed processing in a single run? Email scraperhubapi@gmail.com with a description of what you need.

If this keyword research tool helps your content planning, please leave a review on the actor page. Ratings help other SEO and content teams find it and guide what gets improved next.

---

⚖️ Disclaimer

The Semrush Keyword Magic Tool actor collects keyword suggestions from publicly available Google autocomplete data. It does not access any account, does not require authentication, and does not retrieve content from behind a login.

The metric fields this keyword research tool returns — avgMonthlySearches, lowCPC, highCPC, competitionIndex, competitionValue, intentConfidence, estimatedCTR, contentGapScore, keywordFreshness, monetizationScore and monthlySearchVolumes — are estimates generated deterministically by the actor's own model from the keyword text, country and language. They are not live figures from Semrush, Google Ads, or any third-party keyword data provider, and no affiliation with or endorsement by any such provider is claimed or implied. Verify any figure independently before making a financial or commercial decision based on it.

You are responsible for how you use the output. That includes compliance with the terms of service of any service you query or publish to, and with the laws applicable to you. Keyword data of this kind does not normally contain personal data, but if a keyword you research includes an individual's name or other identifying information, GDPR, UK GDPR, CCPA and equivalent privacy laws may apply to the resulting dataset and you should handle it accordingly.

If you believe data collected by this actor relates to you and should be removed, contact scraperhubapi@gmail.com with the details and the request will be actioned.