Keyboard shortcuts

Press ← or → to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

Tools

The mcp-rss server exposes the following tools.

get_articles

Fetches articles from RSS/Atom feeds.

  • Parameters:
    • feeds (Vec<String>): List of feed URLs to fetch.
    • time_from (Option<String>): Optional ISO 8601 timestamp. Only articles published after this time are returned.
  • Returns: A list of Article objects, deduplicated by URL. Each article has:
    • link (String): Canonical URL of the article.
    • title (String): Article title (empty string if missing).
    • id (String): Feed-specific identifier (GUID).
    • published (Option<String>): Publication date in RFC 3333 format, or null.
    • description (Option<String>): Feed-specific summary/description, or null.
  • Notes: If an article lacks a publication date, it is included regardless of time_from filtering. Articles are deduplicated by link across all feeds. get_articles is never subject to MCP_RSS_ALLOWED_DOMAINS — feeds may be fetched from any host.

fetch_article

Fetches and cleans the text content of a single article from a URL.

  • Parameters:
    • url (String): The URL of the article to fetch.
  • Returns: Cleaned text content (String).
  • Notes: Uses a cascade of CSS selectors (article, main, .post-content, etc.) to extract the main content. Falls back to the full <body> if no specific content block is found.
    • Domain restriction: If the MCP_RSS_ALLOWED_DOMAINS environment variable is set to a comma-separated list of domains, fetch_article only accepts URLs whose hostname is in that list. An entry matches its own host and any of its subdomains (e.g. index.hr also allows www.index.hr). The comparison is case-insensitive and surrounding whitespace around each entry is ignored. When unset, any URL may be fetched. See Restricting article fetching.