About Purili

puri.li — your search, your index.

What is Purili?

Purili is a self-hosted, privacy-first search engine. Instead of relying on a commercial search index, Purili crawls the web itself and builds its own index from scratch. Every result you see comes from pages our own crawler discovered and stored — no third-party search APIs, no tracking, no ads.

How it works

1

Crawl

A Rust-based crawler visits millions of pages, following links up to 4 hops deep from a curated seed list of top domains.

2

Index

Pages are indexed with Tantivy, a full-text search library. Relevance is scored using BM25, the same algorithm behind many production search engines.

3

Serve

Queries hit an axum HTTP API that searches the Tantivy index and returns results in milliseconds. A Next.js frontend renders the results with infocards, tool panels, and bang redirects.

Technology

CrawlerRust + Tokio + redb
Search indexTantivy (BM25)
APIRust + axum
FrontendNext.js + Tailwind
Wiki dataWikipedia XML dump
HostingSelf-hosted VPS

Open source

Purili is open source. You can run your own instance, contribute, or inspect the code. Self-hosting instructions are in the README.

Contact

Reach us at info@sarcque.com. For crawler complaints or bot opt-out, see puri.li/bot.