Open source search database with SQL, HTTP JSON API, full-text, vector, and hybrid search
- Stars11.8k
- Forks628
- Open Issues700
GPL-3.0
- C++
- HTML
- C
About Manticore Search
Manticore Search is an open source database designed for search. It targets applications that need full-text search, filtering, vector retrieval, and stream filtering without Elasticsearch. It is SQL-first and uses the MySQL protocol, so MySQL clients can query it.
Search features include full-text operators, ranking factors, custom ranking, fuzzy search, faceting, geo-spatial search, autocomplete, spelling correction, and highlighting. Hybrid search combines full-text and vector retrieval in one query. It also supports joins, real-time inserts, row-wise and columnar storage, secondary indexes, and a cost-based query optimizer.
Access options include HTTP JSON for data and schema management, with Elasticsearch-compatible writes. Clients are available for PHP, Python, JavaScript, TypeScript, Java, Elixir, Go, and Rust. It includes Galera-based multi-master replication, load balancing, backup tools, SQL BACKUP, S3-compatible backup and restore, and was forked from Sphinx 2.3.2 in 2017.
Key features
- Full-text search with filtering, ranking, faceting, fuzzy search, and highlighting
- Hybrid search combines full-text and vector retrieval in one query
- SQL-first interface with MySQL protocol compatibility
- HTTP JSON API with Elasticsearch-compatible writes
- Built-in multi-master replication, load balancing, and backup tools
Details
- First released
- 2017
- Interfaces
- SQL · MySQL protocol · HTTP JSON
- Search
- Full-text · vector · hybrid
- Storage
- Row-wise · columnar
- Replication
- Multi-master via Galera
- Backup
- manticore-backup · SQL · S3-compatible