MyScale

SQL vector database built on ClickHouse for AI applications

Hand-reviewed by the Toolglade team·Updated Aug 2026

vector-databases#vector-database#sql#clickhouse#rag

Free plan Free trial Claimed API Self-hosted Teams

Toolglade’s take

MyScale's SQL-first approach is a genuine differentiator: teams that live in SQL and need combined metadata plus vector queries will find it more natural than API-driven vector stores. Building on ClickHouse gives it real analytical muscle for filtered and joint queries. The flip side is that a ClickHouse foundation carries operational characteristics that differ from purpose-built vector engines, and it is less hyped than the market leaders. A solid pick for data teams wanting SQL, analytics, and vector search together.

About MyScale

MyScale is an open-source SQL vector database built on ClickHouse that lets you run vector search and filtered joint queries using standard SQL, unifying structured, text, vector, JSON, geospatial, and time-series data.

MyScale (MyScaleDB) is a high-performance SQL vector database built on the open-source ClickHouse engine. Its core idea is that developers should be able to run vector search using the SQL they already know, with vector-related functions available directly in queries, rather than learning a bespoke API for a specialized vector store. Because it is built on ClickHouse, MyScale can manage and process structured data, text, vectors, JSON, geospatial, and time-series data on a single platform, which is especially valuable for filtered searches and joint queries that combine metadata constraints with vector similarity. The team reports better resource efficiency and accuracy for these joint queries compared with integrated options like PostgreSQL with pgvector or Elasticsearch with vector extensions. MyScaleDB went open-source in 2024, and MyScale offers a managed cloud alongside the self-hostable database, appealing to teams that want familiar SQL, strong analytical capabilities, and cost-effective vector search for RAG and other AI workloads.

TL;DR

MyScale is a SQL-first vector database built on ClickHouse, ideal for RAG and analytics teams needing combined metadata and vector queries.

Company overview

MyScale is a company building a SQL-native vector database, MyScaleDB, on top of the open-source ClickHouse engine. MyScaleDB went open-source in 2024, and the company offers a managed cloud service.

It targets teams that want to leverage familiar SQL and ClickHouse's analytical strengths for vector search rather than adopting a specialized vector API.

Product features

MyScale enables vector search through SQL with vector-distance functions, and unifies structured, text, vector, JSON, geospatial, and time-series data. It excels at filtered searches and joint queries combining metadata with similarity.

It integrates with RAG frameworks and is available as both an open-source database and a managed cloud, giving teams flexibility on deployment.