Real-time global intelligence dashboard. AI-powered news aggregation, geopolitical monitoring, and infrastructure tracking in a unified situational awareness interface
MCP server for Naver Search and DataLab APIs
This repository provides an MCP server that exposes Naver search endpoints and DataLab trend endpoints to connected agents. It supports web, news, blog, cafe, image, KnowledgeiN, encyclopedia, local search, category lookup, and shopping or search trend analysis.
Builders who want Claude Code or another MCP-capable agent to query Naver search results and trend data.
You can search Naver and analyze trend data without leaving your agent workflow.
What it does
Naver search tools
Provides tools for web, news, blog, cafe article, image, KnowledgeiN, encyclopedia, and local place search.
DataLab trend analysis
Exposes search-term and shopping trend tools across category, device, gender, age, and keyword views.
Category lookup
Includes `find_category` so you can describe a category in natural language instead of hunting for category numbers.
MCP and skill packaging
Ships as an MCP server and also includes an OpenClaw skill wrapper for installation through ClawHub.
Platform-aware credentials
Supports both NAVER API HUB and legacy Naver Developers keys, with environment variables choosing the platform.
How to get it
- 1This MCP server can also be used from OpenClaw through the ClawHub skill wrapper.…
openclaw skills install naver-search-mcp
- 2The ClawHub skill uses the same published npm package internally
npx -y @isnow890/naver-search-mcp
- 3Run
git clone https://github.com/isnow890/naver-search-mcp.git cd naver-search-mcp npm install npm run build
- 4Navigate to the extracted folder in terminal
cd /path/to/naver-search-mcp npm install npm run build
- 5Finding your path
# Check current location pwd # Absolute path examples # Windows: C:\Users\username\Downloads\naver-search-mcp # macOS: /Users/username/Downloads/naver-search-mcp # Linux: /home/username/Downloads/naver-search-mcp
README
Naver Search MCP Server
MCP server for Naver Search API and DataLab API integration, enabling comprehensive search across various Naver services and data trend analysis.
⚠️ Naver is migrating these APIs — read this before you set up
Naver is moving Search, Search Trend, and Shopping Insight from the Naver Developers Center to NAVER API HUB on NAVER Cloud Platform, and is shutting down three search APIs entirely.
| Date | What happens |
|---|---|
| 2026-07-31 | Developers Center stops accepting new key applications. Shopping / Book / Academic search APIs shut down completely. |
| 2027-06-30 | Developers Center support ends — existing keys stop working. |
Source: Naver Developers Center official notice (Korean) — "Search API, Search Trend, Shopping Insight 서비스 종료 및 NAVER API HUB 이관 안내".
Setting this up for the first time? Get your keys from NAVER API HUB. The Developers Center path closes to new applicants on 2026-07-31, so it is no longer the place to start. Step-by-step instructions are in Option A: NAVER API HUB below.
Already running with Developers Center keys? Nothing to change. They keep working until 2027-06-30, and this version supports both platforms from the same install. When you're ready, get a HUB key and swap the two environment variables — no other changes.
Support plan for this server: both platforms stay supported side by side through Naver's 2027-06-30 cutoff. The Developers Center path will be dropped in a later major version only after that date, once it can no longer work for anyone — so upgrading in the meantime will never take your working setup away.
Three tools were removed in 1.0.49: search_shop, search_book, search_academic. Naver shuts those search APIs down on 2026-07-31 with no replacement on any platform, so there is nothing this server can do to keep them. Shopping Insight (datalab_shopping_*) and find_category are a different API and are not affected.
Available on ClawHub for OpenClaw
You can install this MCP server as an OpenClaw skill from ClawHub:
openclaw skills install naver-search-mcp
The ClawHub skill uses the published npm package internally and needs one credential pair in your OpenClaw environment — either the NAVER API HUB pair (NCP_APIGW_API_KEY_ID / NCP_APIGW_API_KEY) or the Developers Center pair (NAVER_CLIENT_ID / NAVER_CLIENT_SECRET). OpenClaw's apiKey field maps to NAVER_CLIENT_SECRET; HUB users should set both HUB variables explicitly in the environment instead of using apiKey.
Quick Start: Use Without API Key
You can use this server immediately without API keys through Kakao PlayMCP. Simply visit the link and start using it right away!
Tool Details
Available tools:
🆕 Category Search
- find_category: Category search tool so you no longer need to manually check category numbers in URLs for trend and shopping insight searches. Just describe the category in natural language.
Search Tools
- search_webkr: Search Naver web documents
- search_news: Search Naver news
- search_blog: Search Naver blogs
- search_cafearticle: Search Naver cafe articles
- search_image: Search Naver images
- search_kin: Search Naver KnowledgeiN
- search_encyc: Search Naver encyclopedia
- search_local: Search Naver local places
Removed in 1.0.49:
search_shop,search_book,search_academic. Naver shuts down the Shopping / Book / Academic search APIs on 2026-07-31 with no replacement on any platform. This is not a limitation of this server. Shopping Insight (datalab_shopping_*) is a different API and is unaffected.
DataLab Tools
- datalab_search: Analyze search term trends
- datalab_shopping_category: Analyze shopping category trends
- datalab_shopping_by_device: Analyze shopping trends by device
- datalab_shopping_by_gender: Analyze shopping trends by gender
- datalab_shopping_by_age: Analyze shopping trends by age group
- datalab_shopping_keywords: Analyze shopping keyword trends
- datalab_shopping_keyword_by_device: Analyze shopping keyword trends by device
- datalab_shopping_keyword_by_gender: Analyze shopping keyword trends by gender
- datalab_shopping_keyword_by_age: Analyze shopping keyword trends by age group
Getting API Keys
Two platforms issue keys for this server. Set one pair of environment variables — see Configuration below for how the pair you set determines the platform used.
Option A: NAVER API HUB (recommended — the forward path)
- Go to the NAVER Cloud Platform console and sign up or log in.
- Click the region & platform selector in the top right of the console, choose your region/platform, and click Apply.
- Open Menu > All Services > Application Services > NAVER API HUB.
- Click Application, select (or create) your application, then under API management click 인증 정보 (Authentication Info).
- Copy the Client ID and Client Secret shown in the popup.
- Use these as
NCP_APIGW_API_KEY_ID(Client ID) andNCP_APIGW_API_KEY(Client Secret) in the configuration below.
Option B: Naver Developers (legacy — existing keys only)
Naver Developers Center stops accepting new applications on 2026-07-31. If you don't already have a Client ID/Secret from this platform, use NAVER API HUB above instead. Existing Developers Center keys keep working until 2027-06-30.
- Visit Naver Developers and log in with your Naver account
- Click the "Application Registration" (애플리케이션 등록) button
- Fill in the application information:
- Application Name: Enter any name (e.g., "Naver Search MCP")
- Usage: Select "Search" (검색)
- In the API Settings section, check ALL of the following APIs:
- Search (검색) - Required for blog, news, cafe article, web, image, kin, encyclopedia, and local search
- DataLab - Search Trends (데이터랩 - 검색어 트렌드) - Required for search term trend analysis
- DataLab - Shopping Insight (데이터랩 - 쇼핑인사이트) - Required for shopping trend analysis
- Click "Register" to complete registration
- After registration, you'll see your Client ID and Client Secret on the application detail page
- Use these as
NAVER_CLIENT_ID/NAVER_CLIENT_SECRETin the configuration below
Configuration
Credentials
Two platforms are supported. Set one pair — the server picks the platform from which variables you set, since the key strings themselves are indistinguishable.
| Variables | Platform | Endpoint |
|---|---|---|
NCP_APIGW_API_KEY_ID, NCP_APIGW_API_KEY | NAVER API HUB (NCP) | naverapihub.apigw.ntruss.com |
NAVER_CLIENT_ID, NAVER_CLIENT_SECRET | Naver Developers (legacy) | openapi.naver.com |
If both pairs are set, NAVER API HUB wins.
The installation examples below use the legacy NAVER_CLIENT_ID / NAVER_CLIENT_SECRET
pair; swap in the HUB pair from the table above if that's what you have.
Migration to NAVER API HUB
Naver is moving these APIs from the Developers Center to NAVER API HUB on NAVER Cloud Platform.
| Date | What happens |
|---|---|
| 2026-06-25 | NAVER API HUB launched |
| 2026-07-31 | Developers Center stops accepting new applications |
| 2027-06-30 | Developers Center support ends — existing keys stop working |
Existing keys keep working until 2027-06-30. To migrate, get a key from the
NAVER Cloud Platform console and set NCP_APIGW_API_KEY_ID / NCP_APIGW_API_KEY.
Developers Center keys cannot be used against NAVER API HUB.
Installation
Method 1: NPX Installation (Recommended)
The most reliable way to use this MCP server is through NPX. For detailed package information, see the NPM package page.
Claude Desktop Configuration
Add to Claude Desktop config file (%APPDATA%\Claude\claude_desktop_config.json on Windows, ~/Library/Application Support/Claude/claude_desktop_config.json on macOS/Linux):
{
"mcpServers": {
"naver-search": {
"command": "npx",
"args": ["-y", "@isnow890/naver-search-mcp"],
"env": {
"NAVER_CLIENT_ID": "your_client_id",
"NAVER_CLIENT_SECRET": "your_client_secret"
}
}
}
}
Claude Code Configuration
Add to your Claude Code settings:
{
"mcpServers": {
"naver-search": {
"command": "npx",
"args": ["-y", "@isnow890/naver-search-mcp"],
"env": {
"NAVER_CLIENT_ID": "your_client_id",
"NAVER_CLIENT_SECRET": "your_client_secret"
}
}
}
}
Method 2: ClawHub Installation for OpenClaw
This MCP server can also be used from OpenClaw through the ClawHub skill wrapper. Install it with:
openclaw skills install naver-search-mcp
The ClawHub skill uses the same published npm package internally:
npx -y @isnow890/naver-search-mcp
Make sure one credential pair is configured in your OpenClaw environment before using the skill — the NAVER API HUB pair (NCP_APIGW_API_KEY_ID / NCP_APIGW_API_KEY) or the Developers Center pair (NAVER_CLIENT_ID / NAVER_CLIENT_SECRET). OpenClaw's apiKey field maps to NAVER_CLIENT_SECRET; HUB users should set both HUB variables explicitly instead.
Method 3: Local Installation
For local development or custom modifications:
Step 1: Download and Build Source Code
Clone with Git
git clone https://github.com/isnow890/naver-search-mcp.git
cd naver-search-mcp
npm install
npm run build
Or Download ZIP File
- Download the latest version from GitHub Releases
- Extract the ZIP file to your desired location
- Navigate to the extracted folder in terminal:
cd /path/to/naver-search-mcp
npm install
npm run build
⚠️ Important: You must run npm run build after installation to generate the dist folder that contains the compiled JavaScript files.
Step 2: Claude Desktop Configuration
After building, you'll need the following information:
- NAVER_CLIENT_ID: Client ID from Naver Developers
- NAVER_CLIENT_SECRET: Client Secret from Naver Developers
- Installation Path: Absolute path to the downloaded folder
Windows Configuration
Add to Claude Desktop config file (%APPDATA%\Claude\claude_desktop_config.json):
{
"mcpServers": {
"naver-search": {
"type": "stdio",
"command": "cmd",
"args": [
"/c",
"node",
"C:\\path\\to\\naver-search-mcp\\dist\\src\\index.js"
],
"cwd": "C:\\path\\to\\naver-search-mcp",
"env": {
"NAVER_CLIENT_ID": "your-naver-client-id",
"NAVER_CLIENT_SECRET": "your-naver-client-secret"
}
}
}
}
macOS/Linux Configuration
Add to Claude Desktop config file (~/Library/Application Support/Claude/claude_desktop_config.json):
{
"mcpServers": {
"naver-search": {
"type": "stdio",
"command": "node",
"args": ["/path/to/naver-search-mcp/dist/src/index.js"],
"cwd": "/path/to/naver-search-mcp",
"env": {
"NAVER_CLIENT_ID": "your-naver-client-id",
"NAVER_CLIENT_SECRET": "your-naver-client-secret"
}
}
}
}
Path Configuration Important Notes
⚠️ Important: You must change the following paths in the above configuration to your actual installation paths:
- Windows: Change
C:\\path\\to\\naver-search-mcpto your actual downloaded folder path - macOS/Linux: Change
/path/to/naver-search-mcpto your actual downloaded folder path - Build Path: Make sure the path points to
dist/src/index.js(not justindex.js)
Finding your path:
# Check current location
pwd
# Absolute path examples
# Windows: C:\Users\username\Downloads\naver-search-mcp
# macOS: /Users/username/Downloads/naver-search-mcp
# Linux: /home/username/Downloads/naver-search-mcp
Step 3: Restart Claude Desktop
After completing the configuration, completely close and restart Claude Desktop to activate the Naver Search MCP server.
Prerequisites
- A credential pair for one platform — NAVER API HUB or Naver Developers (see Getting API Keys)
- Node.js 18 or higher
- NPM 8 or higher
License
MIT License
Version History
1.0.49 (2026-07-26)
- NAVER API HUB support: set
NCP_APIGW_API_KEY_ID/NCP_APIGW_API_KEYto call the new platform. ExistingNAVER_CLIENT_ID/NAVER_CLIENT_SECRETkeep working unchanged — the platform is chosen by which variable pair you set - Removed
search_shop,search_book,search_academic. Naver shuts those search APIs down on 2026-07-31 with no replacement on any platform - API errors now throw a plain
Errorinstead of the rawAxiosError. Code branching onerror.response?.statusseesundefined; the same detail is inError.message - Error messages name the platform and HTTP status, and 401s suggest checking whether a key was set on the other platform's variables
1.0.48 (2026-05-12)
- Fixed
find_categorydata loading when the MCP server is launched bynpxfrom a different working directory - Category data is now resolved relative to the installed package before falling back to local development paths
1.0.47 (2025-01-03)
- Added "today" keyword support for all DataLab date parameters - no need to call separate time tool
- Fixed server termination issue - MCP server now properly exits when client disconnects
- Added graceful shutdown handlers for SIGINT, SIGTERM, and transport close events
- Removed get_current_korean_time tool - redundant with new "today" keyword feature
- Removed memory monitoring module - resolved setInterval blocking process exit
- Special thanks to @gloomyrobot for reporting the server termination issue
1.0.45 (2025-09-28)
- Resolved platform compatibility issues for hosted MCP installations
- Replaced the Excel export in category search with JSON for better compatibility
- Restored the
search_webkrtool for Korean web search - Improved hosted platform installation compatibility
1.0.44 (2025-08-31)
- Added the
get_current_korean_timetool for essential Korea Standard Time context - Referenced the time tool across existing tool descriptions for temporal queries
- Improved handling of "today", "now", and "current" searches with temporal context
- Expanded Korean date and time formatting outputs with multiple formats
1.0.40 (2025-08-21)
- Added the
find_categorytool with fuzzy matching so you no longer need to check category numbers manually in URLs - Enhanced parameter validation with Zod schema
- Improved the category search workflow
- Implemented a level-based category ranking system that prioritizes top-level categories
1.0.30 (2025-08-04)
- MCP SDK upgraded to 1.17.1
- Fixed compatibility issues with hosted MCP platform specification changes
- Added comprehensive DataLab shopping category code documentation
1.0.2 (2025-04-26)
- README updated: cafe article search tool and version history section improved
1.0.1 (2025-04-26)
- Cafe article search feature added
- Shopping category info added to zod
- Source code refactored
1.0.0 (2025-04-08)
- Initial release
Files in the repo
- data
- scripts
- skills
- src
- tests
- .dockerignore
- .gitignore
- Dockerfile
- LICENSE
- package-lock.json
- package.json
- README-ko.md
- README.md
- RELEASE_NOTES.md
- tsconfig.json
Discussion (0)
Ask about usage, or say what you built with itSign in to join the discussion.
No comments yet. Be the first to say what this is good for.
More connectors
High-performance code intelligence MCP server. Indexes codebases into a persistent knowledge graph — average repo in milliseconds. 158 languages, sub-ms queries, 99% fewer tokens. Single static binary, zero dependencies.

Universal provider proxy for OpenAI Codex & Claude Code — use any LLM (Claude, Gemini, Grok, DeepSeek, Ollama…) with Codex CLI, App, SDK, and Claude Code
Git-native persistent memory for AI coding agents. Implements Google OKF v0.2 with sub-300µs in-memory BM25 search, embedded MCP server, and progressive disclosure. Slashes token bloat by 80% with zero external databases or dependencies. Built in pure Go.
Local-first code intelligence graph for MCP and CLI. Builds a persistent map of your codebase so AI coding tools read only what matters, with benchmarked context reductions on reviews and large-repo workflows.
Stop your AI from making things up — it proposes, deterministic tools decide, every claim checked against ground truth with evidence. Grounded facts and context survive resets. Reverse engineering is the proving ground. MCP server + CLI.