Summary
One blog, three questions, deployed once and shared.
It is scoped to a single blog chosen at deploy time, which makes it a good fit for putting your own archive in front of an assistant rather than searching Hatena at large. Because it runs as a hosted endpoint, several people can point at the same deployment.
What it is
A small service that exposes a single Hatena Blog as three tools — keyword search, recent posts, and one post by URL. It is deployed as a serverless function rather than run locally.
What you get
- search_blog — search that blog's posts by keyword, with a result count you set
- get_recent_posts — the most recent posts, with a result count you set
- get_post_by_url — the full detail of a single post from its URL
- Feed parsing and an in-memory cache, so a run of related questions does not refetch the feed each time
Requirements
The blog id you want to serve, set when the function is deployed, and somewhere to deploy it. Clients reach it over HTTP rather than launching it as a local process.