Skip to content
tatami

Getting started

Install tatami and turn a crawl shard into a searchable columnar file in under a minute.

Three short pages: how tatami thinks about storing a corpus and its search index in one file, how to install the binary and the Go library, and a guided first run that converts a Parquet shard, looks inside it, and reads rows back out.

Introduction Why a crawl corpus and its search index belong in one file, and how tatami puts them there. Installation Install the tatami CLI from a package manager, a prebuilt binary, Go, or the container image, and add the library to a Go module. Quick start Convert a crawl shard to tatami, inspect its layout, read rows back as JSONL, and see the size it saves, in one short session.