Repolode
Submit
quickwit-oss

quickwit-oss/tantivy

A full-text search engine library in Rust, inspired by Apache Lucene, for building custom search into applications.

Open-source alternative to:ElasticsearchApache Solr

View on GitHub
Stars
16,247
Forks
1,017
Open issues
484
Last commit
22 hours ago

What is it?

Tantivy is a Rust crate for building full-text search into your own software. It is not a standalone search server; instead, developers use its API to define schemas, index documents, and run queries. Its design follows Apache Lucene closely, and it uses BM25 scoring, which matches Lucene's ranking.

The library covers the usual search building blocks: configurable tokenizers with stemming for 17 Latin-script languages, phrase and range queries, a natural query language with AND/OR operators, and faceted search. Fields can be text, integers, floats, dates, IP addresses, booleans, or JSON, and fast fields work like Lucene doc values. Documents can be stored with LZ4, Zstd, or no compression, and an aggregation collector provides histograms, range buckets and basic statistics.

Tantivy supports incremental indexing and multithreaded indexing, and it starts in under 10 ms, which suits command-line tools. For Chinese and Japanese tokenization, the project links to third-party tokenizer crates. It runs on stable Rust across Linux, macOS, and Windows.

Key features

  • 01BM25 scoring, the same ranking as Lucene
  • 02Configurable tokenizers with stemming for 17 Latin-script languages
  • 03Phrase, range, and natural-language boolean queries
  • 04Incremental and multithreaded indexing
  • 05Fast fields, faceted search, and aggregations
  • 06Compressed document store with LZ4, Zstd, or no compression

Use cases

  • Adding full-text search to a Rust application without running a separate server
  • Building a custom command-line search tool with fast startup
  • Indexing email or message archives for local search
  • Embedding a search engine in a service that needs Lucene-style querying

Quick start

  1. 1

    Clone the repository

    git clone https://github.com/quickwit-oss/tantivy.git
  2. 2

    Enter the project directory

    cd tantivy
  3. 3

    Run the tests with Cargo (stable Rust)

    cargo test

Things to know

  • It is a Rust library, not a ready-to-run search server, so you must build your own search application around it.
  • Distributed search is out of scope; use Quickwit for that.
  • Documents are immutable, so editing one means deleting and reindexing it.
  • Chinese and Japanese tokenization relies on third-party crates.

Topics

#search-engine#full-text-search#rust#lucene#indexing#bm25

FAQ

What is tantivy?
A full-text search engine library in Rust, inspired by Apache Lucene, for building custom search into applications. Tantivy is a Rust crate for building full-text search into your own software. It is not a standalone search server; instead, developers use its API to define schemas, index documents, and run queries. Its design follows Apache Lucene closely, and it uses BM25 scoring, which matches Lucene's ranking.
Is tantivy free to use? What license does it have?
tantivy is open source under the MIT license. Check the license terms before using it in commercial products.
Is tantivy still maintained?
Actively maintained. The last commit to tantivy was on Oct 21, 2026.
What can I use tantivy for?
Adding full-text search to a Rust application without running a separate server Building a custom command-line search tool with fast startup Indexing email or message archives for local search Embedding a search engine in a service that needs Lucene-style querying
How do I install tantivy?
Clone the repository: git clone https://github.com/quickwit-oss/tantivy.git Enter the project directory: cd tantivy Run the tests with Cargo (stable Rust): cargo test
Where can I get tantivy?
The source code and installation instructions are on GitHub: https://github.com/quickwit-oss/tantivy

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects