MCP server for Korean keyword analysis using Naver SearchAd API. Niche scoring, search volume, trend analysis, blog competition, and batch comparison.
Korean-keyword-mcp has well-structured tool definitions with complete schemas and descriptions. All 6 tools follow verb_noun naming (keyword_expand, niche_score, search_volume, trend, blog_competition, batch_analyze). Input schemas use Zod with proper type constraints (minLength, maxLength, min, max). Descriptions are substantive (100-200 chars), explaining WHAT each tool does and WHEN to use it. However, output schemas are not formally documented, responses are JSON-stringified text without declared field types. Error handling is minimal; failures return generic JSON errors without recovery guidance. No tool annotations (readOnlyHint, idempotentHint) despite all being read-only operations. Parameter descriptions are good but could be more prescriptive about format/constraints.
Batch niche analysis for up to 10 keywords. Calculates full niche score for each keyword in parallel and returns a sorted comparison table. Useful for comparing multiple keyword candidates at once.
Analyze blog competition for a keyword using Naver Search API. Returns total blog post count and top 10 posts with metadata (title, blogger, date).
Expand seed keyword into 50+ related keywords with Quick Score, and calculate Full Niche Score for top N candidates. Returns niche opportunities ranked by score.
Calculate full niche score for a single keyword. Analyzes search volume (Naver SearchAd), blog competition, and 12-month trend to produce a 0-100 score with A-F grade.
Get monthly search volume from Naver SearchAd API. Returns PC/mobile breakdown, competition index (CPC level), and top 20 related keywords with their volumes.
Output schemas not formally documented. All tools return JSON-stringified text without declared field types, making it impossible for LLMs to plan downstream operations or extract specific fields reliably.
No tool annotations despite all tools being read-only. Missing readOnlyHint annotations prevent LLMs from understanding which tools are safe to call speculatively without side effects.
Error handling lacks recovery guidance. When API calls fail (e.g., 'No related keywords found'), responses return generic JSON errors without suggesting next steps or alternative actions.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 62 | 2026-07-28+ | v2 |
Get 12-month search trend from Naver DataLab. Returns monthly relative values (0-100) showing how search interest changed over time.
Parameter descriptions lack format constraints. 'Keyword to analyze' does not explain that keywords are Korean or English, or that the 1-40 char limit is enforced. LLMs cannot self-correct invalid input.
No pagination or result limits documented for related keywords. keyword_expand returns 50+ keywords and batch_analyze accepts up to 10 keywords, but no guidance on handling large result sets or token overhead.