# myscale/MyScaleDB

A @ClickHouse fork that supports high-performance vector search and full-text search.

Repository: https://github.com/myscale/MyScaleDB
Canonical: https://ross.abutalabs.com/products/myscaledb
Homepage: https://myscale.com
Language: C++
License: Apache-2.0
License Family: permissive
Topics: embedding, search-engine, sql, vector-search, vectordb, myscaledb, sql-vector, big-data, llm, rag, ann, similarity-search, image-search, unstructured-analytics
Last push: 2025-02-05T05:11:40+00:00

## Health v2 (maintenance only)
Score: 25/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 5, release rhythm 28, longevity 64
- inputs: {"age_days": 902, "days_push": 574, "days_rel": 682, "gap_med": 40, "n_releases_24m": 2}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1039, forks 72 (observed 2026-08-28T04:03:20.066028+00:00)

## What it is
MyScaleDB is a SQL vector database built as a fork of ClickHouse, adding high-performance vector search and full-text search to a proven OLAP engine. It lets developers build scalable AI and RAG applications using familiar SQL with vector functions, metadata filtering, and SQL-vector joins.

## Use cases
- store and query embeddings with SQL
- build a RAG pipeline with metadata filtering
- run filtered vector search at scale
- combine full-text and vector search in one database
- perform similarity search over images or text
- manage structured and unstructured data for AI apps

## When to choose
- you already know SQL and ClickHouse and want vector search without a new tool
- you need high-precision filtered vector search for RAG
- you want to unify structured, JSON, time-series, and vector data in one OLAP engine
- you need cost-effective scaling for large embedding datasets

## When to avoid
- you need a lightweight embedded vector store for small prototypes
- you require a non-SQL API or ecosystem-specific vector DB features
- you need a fully managed service and don't want to self-host (though MyScale Cloud exists)
- your workload is purely transactional (OLTP) rather than analytical

## Facets
- artifact type: application
- maturity: active
- function: vector-database, database, search-engine
- domain: databases, large-language-models, big-data
- platform: self-hosted, cloud, cpp
- tags: vector-search, clickhouse-fork, sql-vector, similarity-search, rag, embeddings, full-text-search, olap, sql, search, retrieval-augmented-generation, linux, docker

## Member repositories
- myscale/MyScaleDB (main) score 25

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:20.066028+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T07:03:43.557010+00:00, confidence not recorded.
  - readme: https://github.com/myscale/MyScaleDB (fetched 2026-08-28T04:03:20.066028+00:00, sha 3aa41a7ddb05)
  - homepage: https://myscale.com (fetched 2026-08-29T13:04:40.661014+00:00, sha ace7b677121f)
- Data as of 2026-08-30T08:39:29.467469+00:00.