MCP.so
ログイン

Unread MCP Server

@trieloff

Unread MCP Server について

MCP server for searching Unread RSS reader database

基本情報

カテゴリ

その他

ライセンス

Apache-2.0 license

ランタイム

shell

トランスポート

stdio

公開者

trieloff

設定

標準の設定はありません

このサーバーの README には解析可能な MCP 設定ブロックが含まれていません。インストール手順はリポジトリをご確認ください。

リポジトリ

ツール

ツールは検出されませんでした

ツールは README から自動的に抽出されます。メンテナーは ## Tools という見出しの下に記載することで、このタブに反映できます。

概要

What is Unread MCP Server?

A read-only Model Context Protocol (MCP) server for performing keyword-based fulltext searches across articles stored in the Unread RSS reader app.

How to use Unread MCP Server?

Install by cloning the repository to ~/Developer/unread-mcp/ and making unread-mcp.sh executable. Then add the server to your Claude Desktop configuration (claude_desktop_config.json) with the command pointing to that script. Once configured, use Claude to call tools such as search-articles, get-article, get-stats, list-feeds, and search-by-feed.

Key features of Unread MCP Server

  • Keyword-based fulltext search (not semantic/vector)
  • Search across titles, content, authors, and feed names
  • Filter by article status: starred, read, unread
  • Boolean operators (AND, OR, NOT) and exact phrase search
  • Retrieve full article text and database statistics
  • List all RSS feeds with article counts

Use cases of Unread MCP Server

  • Search starred articles for specific topics using Boolean logic
  • Find unread articles about a subject, excluding certain keywords
  • Retrieve the full text of a specific article by its ID
  • View database statistics and list feeds with article counts

FAQ from Unread MCP Server

Is the server read-only?

Yes, the server is strictly read-only and cannot modify your database.

Does it use semantic/vector search?

No, search is keyword-based, not semantic. Use specific terms rather than concepts.

Where does the data live?

The database path is hardcoded to the default Unread database location. The server only reads from that local database.

What dependencies are required?

The server requires bash, sqlite3, and jq. All are standard on macOS.

How can I debug issues?

Check the requests.log file in the same directory as the script for debugging information.

コメント

「その他」の他のコンテンツ