llms.txt
Compare with other sites' llms.txt files →# SereneDB
> SereneDB is a real-time search analytics database that unifies search and analytics behind a PostgreSQL-compatible frontend.
SereneDB combines full-text, vector, and hybrid search with SQL execution, columnar analytics, search indexes, remote tables, and durable RocksDB-backed updates. It is intended for workloads such as search-heavy analytics, AI knowledge bases, logging, internal tools, and product search.
## Primary Pages
- [Homepage](https://serenedb.com/): product overview, architecture summary, installation entry points, and community links.
- [About](https://serenedb.com/about): company and team information.
- [For Press](https://serenedb.com/press): press releases, logos, founder photos, and company boilerplate.
- [Download](https://serenedb.com/download): installation options, packages, Docker entry points, and GitHub releases.
- [Search Benchmark Game](https://serenedb.com/search-benchmark-game): interactive IResearch benchmark results and methodology links.
## Use Cases
- [All use cases](https://serenedb.com/use-cases): the twelve workloads SereneDB serves, grouped into search, analytics and data, and AI and agents, with the one engine they share and links to every use-case page.
- [Hybrid search database](https://serenedb.com/use-cases/hybrid-search-database): BM25 and vector search in one SQL query, filtered ANN, Reciprocal Rank Fusion, and a step-by-step build.
- [RAG database](https://serenedb.com/use-cases/rag-database): the retrieval layer for RAG, with text, vectors, and metadata in one index, hybrid retrieval, LangChain integration, and a step-by-step pipeline.
- [Search over a data lake](https://serenedb.com/use-cases/search-over-data-lake): full-text search over Parquet files in object storage, indexed in place through views over remote files, with a step-by-step S3 setup.
- [Documentation search](https://serenedb.com/use-cases/documentation-search): self-hosted docs search with Serene Docs Search over Git, folder, website, or S3 sources, in full-text, hybrid, or AI-answer mode.
- [Vector search database](https://serenedb.com/use-cases/vector-search-database): vector search in SQL on FLOAT[N] embedding columns beside the rows they describe, embeddings from SQL or your own, HNSW and IVF indexes, kNN, radius and filtered queries, joins and aggregates over the neighbors, and a five-step walkthrough.
- [Full-text search database](https://serenedb.com/use-cases/full-text-search-database): full-text search in SQL on an inverted index, text search dictionaries, stemmed, phrase, prefix and typo-tolerant matching with @@, BM25 and ten other scorers, highlighted snippets, and a five-step walkthrough.
- [OLAP database](https://serenedb.com/use-cases/olap-database): what an OLAP database is and how SereneDB runs one: any Postgres client, a cost-based optimizer, columnar row groups that skip data, ROLLUP, CUBE, PIVOT, QUALIFY, ASOF joins and window functions in SQL, results to clients or files, and a five-step walkthrough.
- [Postgres full text search](https://serenedb.com/use-cases/postgres-full-text-search): Postgres full-text queries from psql or your driver ranked by BM25, indexes over the Postgres you already run, ts_rank vs BM25 on the same rows, typo search, SearchBench against the Postgres family, and a five-step walkthrough.
- [Real-time analytics database](https://serenedb.com/use-cases/real-time-analytics-database): rows visible on commit and searchable at the next index refresh (1000 ms by default), BM25 search and GROUP BY over fresh rows in one SQL statement, ingest over COPY, OpenTelemetry and a subscription to a Postgres publication, Grafana panels, CREATE JOB rollups and retention, and a five-step walkthrough tested on SereneDB 26.11.2.
- [Log analysis](https://serenedb.com/use-cases/log-analysis): OpenTelemetry log records in a search table; phrase, pattern, prefix and trace-id search with @@, GROUP BY on the same rows, OTLP/HTTP, _bulk and read_json ingest, retention jobs, and SearchBench on 1B log rows.
- [Zero-ETL search](https://serenedb.com/use-cases/zero-etl-search): zero ETL by federated query rather than replication, remote PostgreSQL and ClickHouse tables through CREATE SERVER, joins across them in one SQL statement, inverted indexes over remote rows refreshed with REINDEX, a step-by-step Postgres walkthrough, and the limitations.
- [Database for AI agents](https://serenedb.com/use-cases/database-for-ai-agents): one PostgreSQL-protocol login role per agent, column GRANTs checked on every statement, BM25, vector search and aggregates in one query, foreign servers and Parquet views, and the built-in MCP endpoint with check_sql.
## Comparisons
- [All comparisons](https://serenedb.com/alternatives): SereneDB side by side with the search engines, analytics databases and vector databases teams run today, with open benchmarks, sourced capability tables and migration steps.
- [SereneDB vs Elasticsearch](https://serenedb.com/alternatives/elasticsearch-alternative): SearchBench results on 1B OpenTelemetry logs, a capability table, Query DSL next to SQL, the Elasticsearch-compatible API subset and migration steps.
- [SereneDB vs OpenSearch](https://serenedb.com/alternatives/opensearch-alternative): SearchBench results on 1B OpenTelemetry logs against OpenSearch 3.8, a capability table, Query DSL and PPL next to SQL, the Elasticsearch-compatible API subset that OpenSearch clients can use and migration steps.
- [SereneDB vs ClickHouse](https://serenedb.com/alternatives/clickhouse-alternative): SearchBench results on 1B OpenTelemetry logs, BM25 ranking next to ClickHouse's text index, a capability table, ClickHouse SQL next to SereneDB SQL, indexing ClickHouse tables in place and migration steps.
- [SereneDB vs Qdrant](https://serenedb.com/alternatives/qdrant-alternative): VectorBench results on 50M vectors against Qdrant, a capability table, Qdrant's Query API next to SQL for filtered, hybrid, quantized and aggregated vector search and migration steps.
- [SereneDB vs Pinecone](https://serenedb.com/alternatives/pinecone-alternative): vector and hybrid search against Pinecone, a capability table sourced from both vendors' documentation, Pinecone's Python SDK next to SQL for filtered, hybrid, joined and upserted vector search and migration steps.
- [SereneDB vs Tiger Data](https://serenedb.com/alternatives/tigerdata-alternative): SearchBench results on 1B OpenTelemetry logs against TimescaleDB with pg_textsearch, BM25 search and time_bucket analytics in one SQL query, a capability table, TimescaleDB SQL next to SereneDB SQL, copying hypertables and migration steps.
## Markdown Mirrors
- [Homepage markdown](https://serenedb.com/index.md): concise machine-readable product summary.
- [About markdown](https://serenedb.com/about.md): company, team, and architecture summary.
- [For Press markdown](https://serenedb.com/press.md): press materials and media assets.
- [Download markdown](https://serenedb.com/download.md): installation and release links.
- [Benchmark markdown](https://serenedb.com/search-benchmark-game.md): benchmark summary and methodology.
## Technical Documentation
- [Getting started](https://serenedb.com/docs/quick-start): first-run instructions.
- [Installation](https://serenedb.com/docs/installation): Docker and Linux installation guides.
- [SQL reference](https://serenedb.com/docs/sql/introduction): first page of the SQL section, which covers statements, query syntax, functions, data types, and indexes.
- [Configuration](https://serenedb.com/docs/configuration/overview): configuration options, pragmas, limits, logging, and cloud credentials.
- [Search cookbook](https://serenedb.com/docs/cookbook/search): search recipes such as ranking, hybrid search, fuzzy search, facets, and highlighting.
- [Data import and export](https://serenedb.com/docs/data_import_and_export/overview): loading and writing CSV, JSON, and Parquet data.
- [Benchmarks documentation](https://serenedb.com/docs/benchmarks/search-benchmark-game): benchmark details and related analysis.
## Blog And Research
- [SereneDB blog](https://serenedb.com/blog): product updates and engineering articles.
- [Search benchmark overview](https://serenedb.com/blog/search-benchmark-game-overview): context for the Search Benchmark Game results.
- [Vector and hybrid search article](https://serenedb.com/blog/hybrid-search): vector similarity, hybrid search, and embedding management under one index.
- [One SQL, three access modes](https://serenedb.com/blog/one-sql-three-access-modes): local, remote, and federated data access examples.
## Developer Resources
- [GitHub repository](https://github.com/serenedb/serenedb): source code, issues, releases, and contribution entry point.
- [GitHub releases](https://github.com/serenedb/serenedb/releases): downloadable SereneDB and Serene UI builds.
- [Search benchmark repository](https://github.com/serenedb/search-benchmark-game): benchmark harness and methodology.
## Contact And Legal
- [Privacy policy](https://serenedb.com/privacy)
- [Impressum](https://serenedb.com/impressum)
- [Contributor License Agreement](https://serenedb.com/cla)
- Contact email: info@serenedb.com