MyScale is a SQL-compatible vector database that enables developers to build production-grade GenAI applications using familiar SQL queries. It combines vector search, full-text search, and metadata filtering in a single platform, with a proprietary MSTG index engine that delivers 3x faster performance at 3x lower cost. The platform integrates with popular AI frameworks like LangChain and supports RAG workflows, recommendations, and image search use cases.
Funding
Funding not disclosed
Founders
Product
Problem
Building production-grade GenAI applications typically requires developers to learn specialized vector database query languages and manage multiple systems for vector search, full-text search, and metadata filtering. This fragmentation increases development complexity, slows time-to-market, and raises infrastructure costs for teams that already struggle with the learning curve of non-SQL vector databases.
Solution
MyScale provides a fully SQL-compatible vector database that lets developers manage and query AI-related data using standard SQL, eliminating the need to learn proprietary query languages. The platform combines vector search, full-text search with BM25 ranking, and advanced metadata filtering within a single system, enabling complex SQL-vector join queries and text-to-SQL translation. MyScale's proprietary MSTG vector engine delivers 3x faster query performance and 3x lower storage costs compared to alternatives, while supporting configurable tokenizers for multilingual full-text search. The platform integrates deeply with popular AI development frameworks and language SDKs, and includes telemetry features for LLM application observability, making it suitable for RAG pipelines, recommendation systems, chatbots, and image search applications.
Target Audience
Primary customers are developers and engineering teams building GenAI applications, particularly those working on RAG systems, recommendation engines, chatbots, and image search who prefer SQL over specialized vector database query languages.
Features
- Fully SQL-compatible query interface supporting vector search, text search, filtered search, and SQL-vector join operations
- Proprietary MSTG vector index engine delivering 3x faster performance and 3x cost savings
- BM25-based full-text search with configurable tokenizers for multiple languages
- Advanced metadata filtering for precise retrieval in RAG workflows
- Text-to-SQL translation capability for complex query generation
- Deep integrations with AI frameworks including LangChain, plus language SDKs and file format support
- SQL-based role-based access control with SOC 2 and ISO 27001 compliance
- Built-in telemetry for LLM application observability and monitoring