Getting started
Install tsumugi and go from a crawl export to a live, queryable search endpoint in a few minutes.
Three short pages: how tsumugi thinks about a search engine (one file per shard, a cascade over many of them), how to install the binary, and a guided first run that ends with a real search endpoint answering ranked queries over a collection you built.
Introduction
Why a search engine fits in one file per shard, and how tsumugi retrieves and ranks without a transformer on the query path.
Installation
Install tsumugi from Go, Homebrew, Scoop, a release archive, a Linux package, or the container image.
Quick start
From a crawl export to a live search endpoint answering ranked queries, in four commands.