Manticore Search is a high-performance, multi-storage database purpose-built for search and analytics, offering lightning-fast full-text search, real-time indexing, and advanced features like vector search and columnar storage for efficient data analysis. Designed to handle both small and large datasets, it delivers seamless scalability and powerful insights for modern applications.

As an open-source database (available on GitHub), Manticore Search was created in 2017 as a continuation of Sphinx Search engine. Our development team took all the best features of Sphinx and significantly improved its functionality, fixing hundreds of bugs along the way (as detailed in our Changelog). Manticore Search is a modern, fast, and lightweight database with exceptional full-text search capabilities, built on a nearly complete rewrite of its predecessor.

• Query autocomplete
• Fuzzy search
• Over 20 full-text operators and over 20 ranking factors
• Custom ranking
• Stemming
• Lemmatization
• Stopwords
• Synonyms
• Wordforms
• Advanced tokenization at character and word level
• Proper Chinese segmentation
• Text highlighting

Manticore Search supports the ability to add embeddings generated by your Machine Learning models to each document, and then doing a nearest-neighbor search on them. This lets you build features like similarity search, recommendations, semantic search, conversational search, and relevance ranking based on NLP algorithms, among others, including image, video, and sound searches.

Manticore Search supports conversational search over existing vectorized tables. It retrieves the most relevant rows with KNN search, uses those rows together with conversation history as context for the LLM, and can answer through SQL CALL CHAT or the HTTP JSON /search endpoint.

Manticore Search supports JOIN queries via SQL and JSON, allowing you to combine data from multiple tables.

Manticore Search utilizes a smart query parallelization to lower response time and fully utilize all CPU cores when needed.

The cost-based query optimizer uses statistical data about the indexed data to evaluate the relative costs of different execution plans for a given query. This allows the optimizer to determine the most efficient plan for retrieving the desired results, taking into account factors such as the size of the indexed data, the complexity of the query, and the available resources.

Manticore offers both row-wise and column-oriented storage options to accommodate datasets of various sizes. The traditional and default row-wise storage option is available for datasets of all sizes - small, medium, and large, while the columnar storage option is provided through the Manticore Columnar Library for even larger datasets. The key difference between these storage options is that row-wise storage requires all attributes (excluding full-text fields) to be kept in RAM for optimal performance, while columnar storage does not, thus offering lower RAM consumption, but with a potential for slightly slower performance (as demonstrated by the statistics on https://db-benchmarks.com/).

Manticore Columnar Library uses Piecewise Geometric Model index, which exploits a learned mapping between the indexed keys and their location in memory. The succinctness of this mapping, coupled with a peculiar recursive construction algorithm, makes the PGM-index a data structure that dominates traditional indexes by orders of magnitude in space while still offering the best query and update time performance. Secondary indexes are ON by default for all numeric and string fields and can be enabled for json attributes.

Manticore's native syntax is SQL and it supports SQL over HTTP and MySQL protocol, allowing for connection through popular mysql clients in any programming language.

For a more programmatic approach to managing data and schemas, Manticore provides HTTP JSON protocol, similar to that of Elasticsearch.

Manticore supports type='sharding' tables that transparently distribute reads and writes across multiple physical shards on a single node or a replication cluster, improving write scalability while simplifying routing, failover, and operational management.

Manticore provides built-in authentication and authorization for MySQL, HTTP/HTTPS, distributed remote agents, and replication-related operations, with users, bearer tokens, and fine-grained permissions.

You can execute Elasticsearch-compatible insert and replace JSON queries which enables using Manticore with tools like Logstash (version < 7.13), Filebeat and other tools from the Beats family.

Easily create, update, and delete tables online or through a configuration file.

The Manticore Search daemon is developed in C++, offering fast start times and efficient memory utilization. Low-level optimizations further enhance performance. Another crucial component, called Manticore Buddy, is written in PHP and is utilized for high-level functionality that does not require lightning-fast response times or extremely high processing power. Although contributing to the C++ code may pose a challenge, adding a new SQL/JSON command using Manticore Buddy should be a straightforward process.

Newly added or updated documents can be immediately read.

We offer free interactive courses to make learning effortless.

While Manticore is not fully ACID-compliant, it supports isolated transactions for atomic changes and binary logging for safe writes.

Data can be distributed across servers and data centers with any Manticore Search node acting as both a load balancer and a data node. Manticore implements virtually synchronous multi-master replication using the Galera library, ensuring data consistency across all nodes, preventing data loss, and providing exceptional replication performance.

Manticore is equipped with an external tool manticore-backup, and the BACKUP SQL command to simplify the process of backing up and restoring your data. Alternatively, you can use mysqldump to make logical backups.

The indexer tool and comprehensive configuration syntax of Manticore make it easy to sync data from sources like MySQL, PostgreSQL, ODBC-compatible databases, XML, and CSV.

You can integrate Manticore Search with a MySQL/MariaDB server using the FEDERATED engine or via ProxySQL.

You can use Apache Superset and Grafana to visualize data stored in Manticore. Various MySQL tools can be used to develop Manticore queries interactively, such as HeidiSQL and DBForge.

You can also use Manticore Search with Kibana.

Manticore offers a special table type, the "percolate" table, which allows you to search queries instead of data, making it an efficient tool for filtering full-text data streams. Simply store your queries in the table, process your data stream by sending each batch of documents to Manticore Search, and receive only the results that match your stored queries.

Manticore Search is versatile and can be applied in various scenarios, including:

• Full-Text Search:

  • Ideal for e-commerce platforms, enabling fast and accurate product searches with features like autocomplete and fuzzy search.
  • Perfect for content-heavy websites, allowing users to quickly find relevant articles or documents.

• Data Analytics:

  • Ingest data into Manticore Search using Beats/Logstash, Vector.dev, Fluentbit.
  • Analyze large datasets efficiently using Manticore's columnar storage and OLAP capabilities.
  • Perform complex queries on terabytes of data with minimal latency.
  • Visualize data using Kibana, Grafana, or Apache Superset.

• Conversational and AI Search:

  • Build question-answering and assistant experiences on top of existing vectorized content.
  • Ground answers with KNN retrieval and conversation history using conversational search.

• Faceted Search:

  • Enable users to filter search results by categories, such as price, brand, or date, for a more refined search experience.

• Geo-Spatial Search:

  • Implement location-based searches, such as finding nearby restaurants or stores, with Manticore's geo-spatial capabilities.

• Spell Correction:

  • Automatically correct user typos in search queries to improve search accuracy and user experience.

• Autocomplete:

  • Provide real-time suggestions as users type, enhancing search usability and speed.

• Data Stream Filtering:

  • Use percolate tables to filter and process real-time data streams, such as social media feeds or log data, efficiently.

• Architecture: arm64 or x86_64
• OS: Debian-based (e.g. Debian, Ubuntu, Mint), RHEL-based (e.g. RHEL, CentOS, Alma, Oracle Linux, Amazon Linux), Windows, or MacOS.
• Manticore Columnar Library, which provides columnar storage and secondary indexes, requires a CPU with SSE >= 4.2.
• No specific disk space or RAM requirements are needed. An empty Manticore Search instance only uses around 40MB of RSS RAM.

The manual is arranged to reflect the most likely way you would use Manticore:

• starting with basic information about it and how to install and connect
• essential things like adding documents and running searches
• performance optimization tips, tricks and extending Manticore with the help of plugins and custom functions

Key sections of the manual are marked with 1️⃣, 2️⃣, 3️⃣ etc. in the menu for your convenience since their corresponding functionality is most used. If you are new to Manticore we highly recommend not skipping them.

If you are looking for a quick understanding of how Manticore works in general ⚡ Quick start guide is a good place to start.

Each query example has a little icon 📋 in the top-right corner:

You can use it to copy examples to the clipboard. If the query is an HTTP request it will be copied as a CURL command. You can configure the host/port if you press ⚙️.

We love search and we've made our best to make searching in this manual as convenient as possible. Of course it's backed by Manticore Search. Besides using the search bar which requires opening the manual first there is a very easy way to find something by just opening mnt.cr/your-search-keyword :

There are few things you need to understand about Manticore Search that can help you follow the best practices of using it.

• Real-time table allows adding, updating and deleting documents with immediate availability of the changes.
• Plain table is a mostly immutable data structure and a basic element used by real-time tables. Plain table stores a set of documents, their common dictionary and indexation settings. One real-time table can consist of multiple plain tables (chunks), but besides that Manticore provides direct access to building plain tables using tool indexer. It makes sense when your data is mostly immutable, therefore you don't need a real-time table for that.

Manticore Search works in two modes:

• Real-time mode (RT mode). This is a default one and allows to manage your data schema imperatively:
  • allows managing your data schema online using SQL commands CREATE/ALTER/DROP TABLE and their equivalents in non-SQL clients
  • in the configuration file you need to define only server-related settings including data_dir
• Plain mode allows to define your data schemas in a configuration file, i.e. provides declarative kind of schema management. It makes sense in three cases:
  • when you only deal with plain tables
  • or when your data schema is very stable and you don't need replication (as it's available only in the RT mode)
  • when you have to make your data schema portable (e.g. for easier deployment of it on a new server)

You cannot combine the 2 modes and need to decide which one you want to follow by specifying data_dir in your configuration file (which is the default behaviour). If you are unsure our recommendation is to follow the RT mode as if even you need a plain table you can build it with a separate plain table config and import to your main Manticore instance.

Real-time tables can be used in both RT and plain modes. In the RT mode a real-time table is defined with a CREATE TABLE command, while in the plain mode it is defined in the configuration file. Plain (offline) tables are supported only in the plain mode. Plain tables cannot be created in the RT mode, but existing plain tables made in the plain mode can be converted to real-time tables and imported in the RT mode.

Manticore provides multiple ways and interfaces to manage your schemas and data, but the two main are:

• SQL. This is a native Manticore's language which enables all Manticore's functionality. The best practice is to use SQL to:
  • manage your schemas and do other DBA routines as it's the easiest way to do that
  • design your queries as SQL is much closer to natural language than the JSON DSL which is important when you design something new. You can use Manticore SQL via any MySQL client or /sql.
• JSON. Most functionality is also available via JSON domain specific language. This is especially useful when you integrate Manticore with your application as with JSON you can do it more programmatically than with SQL. The best practice is to first explore how to do something via SQL and then use JSON to integrate it into your application.

Docker images of Manticore Search are publicly accessible on Docker Hub, built from the Manticore Search docker GitHub repository.

To retrieve the Manticore image, run the following command:

docker pull manticoresearch/manticore

For more information about using Manticore in Docker, see the Using Manticore in Docker section.