Keyword data

These APIs retrieve search data from Google and Naver. The fan-out queries an AI issues when building an answer are ultimately search terms. You need to know which search terms have demand and how competitive they are before you can decide what content to create to get cited by AI.

Google keyword metrics

Retrieve search volume, difficulty, CPC, and search intent for a list of keywords (1 to 100, each up to 80 characters) in one call. The example below uses the Korean keywords for “used car” and “used car prices”.

curl -X POST "https://platform-api.trychainshift.ai/api/v1beta/platform/keywords/google/metrics" \
  -H "API-Key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"keywords": ["중고차", "중고차 시세"]}'

Below is a real response (monthly search volume is trimmed to the latest 2 months; the API actually returns up to 7 years).

{
  "languageCode": "ko",
  "countryCode": "KR",
  "items": [
    {
      "keyword": "중고차",
      "metrics": {
        "searchVolume": 74000,
        "keywordDifficulty": 74,
        "cpc": 1.34,
        "searchIntent": "transactional",
        "searchVolumeTrend": { "monthly": 0, "quarterly": 0, "yearly": 0 },
        "monthlySearches": [
          { "year": 2026, "month": 6, "searchVolume": 60500 },
          { "year": 2026, "month": 5, "searchVolume": 60500 }
        ]
      }
    },
    {
      "keyword": "중고차 시세",
      "metrics": {
        "searchVolume": 6600,
        "keywordDifficulty": 18,
        "cpc": 2.11,
        "searchIntent": "commercial",
        "searchVolumeTrend": { "monthly": 0, "quarterly": 22, "yearly": 0 },
        "monthlySearches": [
          { "year": 2026, "month": 6, "searchVolume": 6600 },
          { "year": 2026, "month": 5, "searchVolume": 6600 }
        ]
      }
    }
  ]
}

What these two keywords show: “중고차” (used car) is searched 74,000 times a month but has an SEO difficulty of 74, making it hard to break into, while “중고차 시세” (used car prices) has a search volume of 6,600 with a low difficulty of 18 and quarterly search volume growing 22%. Low-difficulty, growing keywords like this take content priority over the big head terms.

searchIntent: search intent, one of informational (looking for information) / commercial (evaluating a purchase) / transactional (ready to buy) / navigational (looking for a specific site)
keywordDifficulty: SEO entry difficulty (0 to 100)
Change the target market with languageCode (default ko) and countryCode (default KR).

If metrics is null, there is no data for that keyword. If metrics is present but searchVolume is null, the keyword falls into a restricted Google Ads category such as healthcare or finance, and search volume is not disclosed.

Retrieve related keywords (related) or autocomplete long-tail suggestions (suggestions) from a single seed keyword. limit defaults to 30 and is capped at 100.

curl -H "API-Key: YOUR_API_KEY" \
  "https://platform-api.trychainshift.ai/api/v1beta/platform/keywords/google/related?keyword=중고차&limit=5"

Looking only at the related keywords portion of a real response (each keyword’s metrics structure is the same as above):

Related keyword	Search volume	Difficulty	Intent	Yearly trend
중고차 시세	6,600	18	commercial	0%
sk 엔카 중고차	4,400	68	transactional	-34%
중고차 판매 사이트	1,300	78	commercial	-19%
중고차 매매	1,300	47	transactional	-23%

You can also read from this data that the branded search “sk 엔카 중고차” (SK Encar used cars) is declining 34% year over year. Keyword data lets you observe the signal of search shifting to AI answers.

Naver keyword metrics

Naver provides metrics (metrics), related keywords (related), autocomplete suggestions (suggestions), and a 12-month search volume trend (trend). It is specific to the Korean market, so there are no language or country parameters.

Naver metrics and trend accept 1 to 15 keywords per call (Google accepts 1 to 100).

curl -X POST "https://platform-api.trychainshift.ai/api/v1beta/platform/keywords/naver/metrics" \
  -H "API-Key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"keywords": ["중고차"]}'

{
  "items": [
    {
      "keyword": "중고차",
      "metrics": {
        "searchVolumeTotal": 272300,
        "searchVolumePC": 73000,
        "searchVolumeMobile": 199300,
        "pcBelowThreshold": false,
        "mobileBelowThreshold": false,
        "avgClickPC": 4358.3,
        "avgClickMobile": 21333.4,
        "avgCTRPC": 6.44,
        "avgCTRMobile": 11.44,
        "competition": "높음",
        "adDepth": 10
      }
    }
  ]
}

For the same keyword “중고차”, Naver’s monthly search volume (272,300) is 3.7 times Google’s (74,000), and 73% of it is mobile. This is why a content strategy for the Korean market needs to look at Naver data separately.

avgClickPC / avgClickMobile: average monthly clicks; avgCTRPC / avgCTRMobile: average click-through rate (%)
competition: level of ad competition (낮음 low / 중간 medium / 높음 high); adDepth: average number of ads shown
Keywords with very low search volume return pcBelowThreshold / mobileBelowThreshold as true, and exact figures are not provided.

Naver search volume trend

Relative search volume for the last 12 months. The values are not absolute; they are ratios where the maximum in the period is 100. You can convert to absolute values using anchor (the absolute search volume for that month).

curl -X POST "https://platform-api.trychainshift.ai/api/v1beta/platform/keywords/naver/trend" \
  -H "API-Key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"keywords": ["중고차"]}'

Below is a real response (trimmed to 3 of the 12 points).

{
  "items": [
    {
      "keyword": "중고차",
      "trend": {
        "points": [
          { "year": 2025, "month": 7, "ratio": 100 },
          { "year": 2025, "month": 12, "ratio": 75.87 },
          { "year": 2026, "month": 6, "ratio": 65.71 }
        ],
        "anchor": { "year": 2026, "month": 6, "searchVolumeTotal": 272300 }
      }
    }
  ]
}

Naver search volume for “중고차” dropped by more than a third in one year (100 → 65.7). To see where that search demand went, and whether it moved to AI answers, look at it together with your run collection data.

Site keyword analysis

Give it a website URL and it crawls the site to extract core keywords. It follows the same submit → poll pattern as running prompts.

curl -X POST "https://platform-api.trychainshift.ai/api/v1beta/platform/siteKeywordAnalyses" \
  -H "API-Key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url": "https://acme.com", "maxPages": 30}'

{ "taskID": "sK4jN8bV2xC6mZ1qW5eR9t" }

maxPages is required and must be between 1 and 50. Poll the status with GET /siteKeywordAnalyses/{taskID}. The status moves from PENDING → IN_PROGRESS → COMPLETED (or FAILED), and once complete the response includes the extracted keyword list. Use the includeSourceURLs and includeContext parameters to also receive the URLs and context where each keyword was found.

Error cases

Code	HTTP	Cause
`KEYWORD_COUNTRY_NOT_SUPPORTED`	400	Unsupported country code
`KEYWORD_METRICS_PROVIDER_REJECTED`	400	The metrics provider rejected the request (for example, an invalid language and country combination)
`KEYWORD_METRICS_PROVIDER_UNAVAILABLE`	502	The external metrics provider call failed. Retry after a short wait
`KEYWORD_METRICS_NOT_CONFIGURED`	503	The keyword metrics feature is disabled in this environment

Getting started

AI models

Authentication

Guides

MCP

Google keyword metrics

Naver keyword metrics

Naver search volume trend

Site keyword analysis

Error cases

Google keyword metrics

Google related keywords and autocomplete

Naver keyword metrics

Naver search volume trend

Site keyword analysis

Error cases