[![MseeP.ai Security Assessment Badge](https://mseep.net/pr/vectorinstitute-mcp-goodnews-badge.png)](https://mseep.ai/app/vectorinstitute-mcp-goodnews) # MCP Goodnews --- [![CodeQL](https://github.com/VectorInstitute/mcp-goodnews/actions/workflows/github-code-scanning/codeql/badge.svg)](https://github.com/VectorInstitute/mcp-goodnews/actions/workflows/github-code-scanning/codeql) [![Linting](https://github.com/VectorInstitute/mcp-goodnews/actions/workflows/lint.yml/badge.svg)](https://github.com/VectorInstitute/mcp-goodnews/actions/workflows/lint.yml) [![Unit Testing and Upload Coverage](https://github.com/VectorInstitute/mcp-goodnews/actions/workflows/unit_test.yml/badge.svg)](https://github.com/VectorInstitute/mcp-goodnews/actions/workflows/unit_test.yml) [![codecov](https://codecov.io/github/VectorInstitute/mcp-goodnews/graph/badge.svg?token=KvwFM5bQiH)](https://codecov.io/github/VectorInstitute/mcp-goodnews) [![Release](https://github.com/VectorInstitute/mcp-goodnews/actions/workflows/release.yml/badge.svg)](https://github.com/VectorInstitute/mcp-goodnews/actions/workflows/release.yml) ![GitHub License](https://img.shields.io/github/license/VectorInstitute/mcp-goodnews)

MCP Goodnews is a simple Model Context Protocol (MCP) application that features a server for getting good, positive, and uplifting news. This tool fetches news articles from the [NewsAPI](https://newsapi.org/) and uses a Cohere LLM to rank and return the top news articles based on positive sentiment. Read the [blog post](https://medium.com/data-science-collective/goodnews-mcp-good-news-at-your-fingertips-d6cda34d558d) on Medium! ## Motivation In a world where negative news often dominates headlines, Goodnews MCP aims to shine a light on more positive and uplifting news stories. This project was inspired by an earlier initiative called GoodnewsFirst, which delivered positive news daily to email subscribers — it was a really awesome project! While GoodnewsFirst predated recent breakthroughs in Large Language Models (LLMs) and relied on traditional methods for sentiment ranking, Goodnews MCP leverages modern LLMs to perform sentiment analysis in a zero-shot setting. ## Example Usage: MCP Goodnews with Claude Desktop

### Requirements - [Cohere API Key](https://dashboard.cohere.com/) - [NewsAPI Key](https://newsapi.org/) - [Claude Desktop Application](https://claude.ai/download) - [uv Python Project and Package Manager](https://docs.astral.sh/uv/getting-started/installation/) ### Clone `mcp-goodnews` ```bash # Clone the repository git clone https://github.com/VectorInstitute/mcp-goodnews.git ``` In the next step, we'll need to provide the absolute path to the location of this cloned repository. ### Update Claude Desktop Config to find mcp-goodnews #### For Mac/Linux ```bash # Navigate to the configuration directory cd ~/Library/Application\ Support/Claude/config # Edit the claude_desktop_config.json file nano claude_desktop_config.json ``` #### For Windows ```bash # Navigate to the configuration directory cd %APPDATA%\Claude\config # Edit the claude_desktop_config.json file notepad claude_desktop_config.json ``` And you'll want to add an entry under `mcpServers` for `Goodnews`: ```json { "mcpServers": { "Goodnews": { "command": "/uv", "args": [ "--directory", "/mcp-goodnews/src/mcp_goodnews", "run", "server.py" ], "env": { "NEWS_API_KEY": "", "COHERE_API_KEY": "" } } } } ``` ### Start or Restart Claude Desktop Claude Desktop will use the updated config to build and run the mcp-goodnews server. If successful, you will see the hammer tool in the bottom-right corner of the chat dialogue window.

Clicking the hammer tool icon will bring up a modal that lists available MCP tools. You should see `fetch_list_of_goodnews` listed there.

### Ask Claude for Good News Example prompts: - "Show me some good news from today." - "What positive things happened in the world this week?" - "Give me uplifting news stories about science." ## How It Works 1. When you request good news, the application queries the NewsAPI for recent articles 2. The Cohere LLM analyzes the sentiment of each article 3. Articles are ranked based on positive sentiment score 4. The top-ranking good news stories are returned to you through Claude ## License [Apache 2.0](./LICENSE) --- _Stay positive with Goodnews MCP!_