The useful field here is the stars gained in the period rather than the total: it separates a project that is moving now from one that has been large for years. The spoken-language filter is the other half — it surfaces work that never reaches an English-language feed.
A small Python server that fetches GitHub's trending pages and returns them as structured JSON. Two calls: trending repositories and trending developers, each filterable by programming language and by daily, weekly or monthly window.
- Trending repositories with full name, URL, description, language, stars, forks and the stars gained in the period
- Trending developers with their handle, name, profile URL and the repository they are trending for
- Filters on programming language and on the period — daily, weekly or monthly
- A spoken-language filter on repositories, for finding what is trending outside English
- JSON responses, so the model reads fields rather than parsing a page
Python 3.12, then pip install mcp-github-trending. No GitHub token and no account: it reads the public trending pages.
One command — pip install mcp-github-trending
