Skip to main content
Glama
narumiruna

coindesk-mcp

README.md
# Crypto News MCP Server

## Usage

From Docker

```json
{
  "mcpServers": {
    "cryptonewsmcp": {
      "command": "docker",
      "args": ["run", "-i", "--rm", "narumi/crypto-news-mcp"]
    }
  }
}
```

From PyPI

```json
{
  "mcpServers": {
    "cryptonewsmcp": {
      "command": "uvx",
      "args": ["cryptonewsmcp"]
    }
  }
}
```

From GitHub

```json
{
  "mcpServers": {
    "cryptonewsmcp": {
      "command": "uvx",
      "args": [
        "--from",
        "git+https://github.com/narumiruna/crypto-news-mcp",
        "cryptonewsmcp"
      ]
    }
  }
}
```

## Components

### Tools

- recent_news
  - Fetches latest crypto news from specified site
  - Input:
    - site: Source site ("coindesk" or "decrypt")
  - Returns formatted list of news entries with titles, links, dates and summaries
- read_news
  - Fetches article HTML from URL and converts it to Markdown
  - Input:
    - url: Article URL to retrieve
  - Returns Markdown-formatted article content

TDQS

A4.1/5.0

Scored across 2 tools

Disambiguation5/5

The two tools have clearly distinct purposes: read_news extracts full content from a specific article URL, while recent_news fetches a list of recent articles from an RSS feed. There is no overlap in functionality, and an agent can easily distinguish between retrieving a single article versus getting multiple recent articles.

Naming Consistency5/5

Both tools follow a consistent verb_noun pattern (read_news and recent_news) with clear, descriptive names. The naming convention is uniform throughout the tool set, making it predictable and easy for agents to understand.

Tool Count2/5

With only 2 tools, the server feels thin for a news domain. While the tools cover reading specific articles and listing recent ones, there are likely gaps such as searching news, filtering by category, or accessing historical archives. The count is too low for comprehensive news interaction.

Completeness2/5

The tool surface is significantly incomplete for a news server. It lacks essential operations like searching for articles by keyword, filtering by date or category, and accessing older or archived news. Agents will struggle with basic news-related tasks beyond the two provided functions.