A Tiny Search Tool

Goal

Make the notes on this site searchable without adding a backend, a database, or a hosted search service. I wanted something small enough to understand completely and cheap enough to run forever.

Approach

The plan is to generate a single small JSON index at build time — titles, tags, and a plain-text version of each note's content — and search through it entirely in the browser. No network request after the initial page load, no server-side logic at all.

What's working

Basic word matching against the index is done and works well for a site this size. Typing a word highlights matching notes instantly, since the whole index fits comfortably in memory for the number of pages here.

What's still unresolved

Ranking results by relevance rather than just presence is the open problem right now. A simple frequency count is good enough for a few dozen notes, but I'd like something a little smarter before this scales past that.

Notes

I'm deliberately not reaching for an existing search library yet. Part of the point of this project is understanding how much a from-scratch approach can carry before it becomes worth the dependency.