SereneDB

Product Introduction

1. Definition: SereneDB is a real-time search analytics database and open-source SQL database engine available at https://serenedb.com. It is the result of 12 years of development and combines ultra-fast full-text search, vector search, hybrid search, fuzzy search, geospatial search, SQL execution, columnar analytics, durable storage, and a PostgreSQL-compatible frontend in one system. The project is Apache 2.0 licensed, with source code at https://github.com/serenedb/serenedb and builds for Linux, macOS, Windows, and Docker.
2. Core Value Proposition: SereneDB exists to unify search and analytics workloads that traditionally require separate systems, such as Elasticsearch for search and ClickHouse or PostgreSQL extensions for analytics. It is Postgres- and Elastic-compatible, so you can keep your SQL, your drivers, and your Elastic clients while dropping the second system and the ETL between them. In its public benchmark, SereneDB reports outperforming Elasticsearch, ClickHouse, and Postgres search extensions, indexing 1B logs in under 8 minutes at approximately 10x less disk, with Apache 2.0 licensing, public methodology, and raw results.

Main Features

1. Unified Full-Text, Vector, Hybrid, and Geospatial Search: SereneDB integrates full-text, fuzzy, geospatial, vector, and hybrid search directly over analytical data. This means search queries and analytical SQL can run against the same dataset without exporting data to a separate search cluster. The engine combines search indexes with columnar storage and durable updates, enabling real-time search analytics rather than batch-only search or delayed analytics.
2. SQL Execution Powered by Velox: SereneDB executes optimized joins, unions, aggregations, and large-scale analytical queries using Velox. Velox provides a high-performance execution layer for SQL analytics, while the PostgreSQL-compatible frontend lets users connect through the PostgreSQL protocol from familiar tools and clients. This design supports mixed workloads where full-text search results are joined or aggregated with analytical data in SQL.
3. Hybrid Storage with RocksDB-Backed Durable Updates: SereneDB uses hybrid storage that includes search indexes, columnar storage, remote tables, and RocksDB-backed durable updates. Columnar storage accelerates analytical scans and aggregations, search indexes serve full-text and vector queries, remote tables are part of the hybrid storage layer, and RocksDB supports durable updates for changing data. The result is a database that can handle both high-ingest log data and low-latency search analytics.
4. PostgreSQL-Compatible Frontend and Elastic-Compatible Clients: SereneDB provides a PostgreSQL-compatible frontend, so existing SQL, PostgreSQL drivers, and many PostgreSQL tools can connect through the PostgreSQL protocol. It is also Elastic-compatible, allowing teams to keep Elastic clients where applicable. This compatibility reduces migration friction and avoids rewriting application code or replacing client libraries when consolidating search and analytics.
5. Open-Source Apache 2.0 Distribution and Agent-Readable Resources: SereneDB is open source under Apache 2.0, with public methodology and raw benchmark results. Users can install with curl https://install.serenedb.com | sh, download builds from https://serenedb.com/download, or follow the getting started guide at https://docs.serenedb.com/quick-start. Machine-readable resources include https://serenedb.com/llms.txt, https://serenedb.com/index.md, and https://serenedb.com/sitemap.xml for discovery by agents and crawlers.

Problems Solved

1. Pain Point: SereneDB addresses the fragmented search and analytics stack problem: teams often run Elasticsearch for full-text and vector search, ClickHouse or another columnar database for analytics, and PostgreSQL for transactional or SQL workloads. This creates ETL pipelines, duplicate data, synchronization lag, higher infrastructure costs, and operational complexity. It also addresses slow log indexing, high disk usage, and the need to maintain separate query languages, drivers, and clients for search versus analytics.
2. Target Audience: SereneDB is built for data engineers, platform engineers, SREs, DevOps teams, backend developers, observability engineers, security analytics teams, and AI/ML engineers building vector or hybrid search applications. It is also relevant to PostgreSQL users who need search and analytics, Elasticsearch users seeking lower disk footprint or simpler architecture, and CTOs or infrastructure leaders reducing database sprawl and ETL overhead.
3. Use Cases: SereneDB is essential for real-time log analytics, observability, security event investigation, and large-scale SQL analytics where full-text search must be combined with aggregations. It supports e-commerce and enterprise search with analytics, geospatial search over analytical data, vector and hybrid search for AI/RAG applications, and replacement or consolidation of Elasticsearch plus ClickHouse pipelines. Its public benchmark use case indexes 1B logs in under 8 minutes at approximately 10x less disk, making it relevant for high-volume log ingestion and search analytics.

Unique Advantages

1. Differentiation: Unlike traditional approaches that combine Elasticsearch for search, ClickHouse for analytics, and PostgreSQL extensions for SQL search, SereneDB unifies full-text search and fast analytics in one engine. It is Postgres- and Elastic-compatible, so it preserves SQL, drivers, and Elastic clients while removing the second system and the ETL between them. In its public benchmark, it reports outperforming Elasticsearch, ClickHouse, and Postgres search extensions, with Apache 2.0 licensing and public methodology and raw results for verification.
2. Key Innovation: The key innovation is the combination of search indexes, columnar analytics, remote tables, RocksDB-backed durable updates, and Velox-powered SQL execution behind a PostgreSQL-compatible frontend. This hybrid architecture lets one database handle ultra-fast full-text search, vector and hybrid search, geospatial search, and large-scale analytical SQL without separate systems. The 12-year development cycle and open-source Apache 2.0 release make the engine available for evaluation, benchmarking, and production use across Linux, macOS, Windows, and Docker.

Frequently Asked Questions (FAQ)

1. What is SereneDB? SereneDB is an open-source real-time search analytics database that combines full-text, vector, hybrid, fuzzy, and geospatial search with fast SQL analytics and a PostgreSQL-compatible frontend. It is available at https://serenedb.com and licensed under Apache 2.0.
2. Is SereneDB a replacement for Elasticsearch and ClickHouse? SereneDB is designed as a unified alternative to running Elasticsearch for search and ClickHouse for analytics. It is Postgres- and Elastic-compatible, so teams can keep SQL, drivers, and Elastic clients while dropping the second system and the ETL between them. Its public benchmark reports outperforming Elasticsearch, ClickHouse, and Postgres search extensions.
3. Does SereneDB support vector search and hybrid search? Yes. SereneDB supports full-text, fuzzy, geospatial, vector, and hybrid search over analytical data, with search indexes integrated into the same engine as columnar analytics and SQL execution.
4. Is SereneDB PostgreSQL-compatible? Yes. SereneDB provides a PostgreSQL-compatible frontend and connects through the PostgreSQL protocol, allowing familiar PostgreSQL tools, drivers, and SQL workflows to query the database.
5. How fast is SereneDB for log indexing and disk usage? In its public benchmark, SereneDB indexes 1B logs in under 8 minutes at approximately 10x less disk. The benchmark methodology and raw results are public, and SereneDB is licensed under Apache 2.0.