Search blog posts on Naver
Search news articles on Naver
Search books on Naver
Get detailed book information using title or ISBN
Check if a search term is adult content
Search encyclopedia entries on Naver
Search articles in Naver cafes
Search questions and answers on Naver
Search local information on Naver
Correct spelling errors in a given text
Search web pages on Naver
Search images on Naver with filters
Search shopping items on Naver with filters
Search documents on Naver
The Naver API MCP provides a comprehensive interface to Naver's extensive search capabilities. It allows you to search across multiple Naver services including blogs, news, books, images, shopping, and local information. With tools for spelling correction and adult content filtering, it offers a complete solution for Korean-language search needs. This MCP wraps Naver's OpenAPI services in a convenient interface that follows the Model Context Protocol, making it easy to integrate with AI assistants. It requires Naver developer credentials which can be obtained for free from the Naver Developers portal.
The Naver API MCP provides access to Naver's search services through a Model Context Protocol server. This allows AI assistants to search and retrieve information from various Naver platforms including blogs, news, books, shopping, and more.
Before using this MCP, you need to obtain Naver API credentials:
You can install the Naver API MCP using pip:
pip install mcp-naver
Then run the server with your Naver credentials:
python -m mcp-naver.hosts.claude_desktop \
-e NAVER_CLIENT_ID=<YOUR NAVER CLIENT ID> \
-e NAVER_CLIENT_SECRET=<YOUR NAVER CLIENT SECRET>
Alternatively, you can use uv:
uv pip install mcp-naver
uv run python -m mcp-naver.hosts.claude_desktop \
-e NAVER_CLIENT_ID=<YOUR NAVER CLIENT ID> \
-e NAVER_CLIENT_SECRET=<YOUR NAVER CLIENT SECRET>
For Cursor, install and run with:
pip install mcp-naver
python -m mcp-naver.hosts.cursor \
-e NAVER_CLIENT_ID=<YOUR NAVER CLIENT ID> \
-e NAVER_CLIENT_SECRET=<YOUR NAVER CLIENT SECRET>
Or with uv:
uv pip install mcp-naver
uv run python -m mcp-naver.hosts.cursor \
-e NAVER_CLIENT_ID=<YOUR NAVER CLIENT ID> \
-e NAVER_CLIENT_SECRET=<YOUR NAVER CLIENT SECRET>
You can also install from source:
git clone https://github.com/pfldy2850/py-mcp-naver.git
cd py-mcp-naver
uv sync --dev --all-extras
Run it with FastMCP CLI:
fastmcp install mcp_naver/server.py \
-e NAVER_CLIENT_ID=<YOUR NAVER CLIENT ID> \
-e NAVER_CLIENT_SECRET=<YOUR NAVER CLIENT SECRET>
Once installed, the MCP provides various tools for searching Naver services. You can use these tools to search for blogs, news, books, images, and more.
For example, to search for blog posts about "Korean food":
search_blog(query="Korean food", display=5, sort="sim")
To search for local restaurants:
search_local(query="restaurants in Seoul", display=10, sort="random")
Most search functions accept parameters like:
query: The search termdisplay: Number of results to return (default: 10)start: Starting position for pagination (default: 1)sort: Sort order (usually "sim" for relevance or other options like "date")Some specialized functions like image search and shopping search offer additional filters.
The MCP requires two environment variables:
NAVER_CLIENT_ID: Your Naver API Client IDNAVER_CLIENT_SECRET: Your Naver API Client SecretThese can be provided when starting the server as shown in the installation instructions.