Date and category filters exist on every news API. Passing a block of text and getting back related articles is the query that turns a claim you are checking into the coverage around it, without having to guess which keywords the outlet used. The prompts are the other half of the design: the briefing shapes are already written, so a daily summary is one invocation rather than a query you reassemble each morning.
A Python MCP server over the API behind boring-news.fr, a French news site. It covers the four ways you actually look for a story — by when it ran, by who is in it, by what it is about, and by resemblance to something you already have — and ships prompts that turn those into daily briefings.
- Articles for a date, narrowed by category or tags — `get_articles_by_date`
- Articles that mention a named person — `get_articles_by_person`
- Articles similar to a passage of text you supply, rather than to a keyword — `get_similar_articles`
- The day's article groups and the categories available, so a query can be scoped before it is run — `get_article_groups`, `get_categories`
- Four ready-made briefing prompts for a given date: a tech-and-culture daily, a comprehensive summary, a highlights pass and a cultural-news pass — `daily_news`, `daily_news_summary`, `daily_news_highlights`, `daily_cultural_news`
No account and no key — it reads the public Boring News API. Python >= 3.8, plus httpx >= 0.25.0 and fastmcp >= 0.1.0, which come with the package. Install with `pip install boring-news-mcp`; the package is `boring-news-mcp` (0.5.2 in pyproject) and runs over stdio as a Python module. The content is French-language news, so this is a source for French coverage rather than a general news feed.
One command — pip install boring-news-mcp