Cottontail DB

Category: Vector Database Engines
Tags: open-source, vector-databases, high-dimensional, vector-search

Description

Cottontail DB is an open-source column store vector database designed for multimedia retrieval. It supports both classical boolean retrieval and vector-space (nearest neighbor) search for similarity queries, utilizing a unified data and query model. The system is intended for use in research and production environments involving high-dimensional data.

Features

• Column store architecture optimized for multimedia retrieval.
• Support for classical boolean queries.
• Support for vector-space (nearest neighbor) queries for similarity search.
• Unified data and query model for both boolean and vector-space retrieval.
• Communication via gRPC (default port 1865) for easy integration.
• Pre-built stubs and client API available for Kotlin and Java via Maven dependency.
• Command-line interface (CLI) bundled with the distribution; can also be built separately.
• Available as a Docker image from DockerHub for easy deployment.
• Requires Java 11 or newer (OpenJDK or Oracle JDK).
• Gradle-based build system with wrapper included for out-of-the-box setup.
• Compatibility with vitrivr components.
• Comprehensive documentation and example repositories available.

Pricing

Cottontail DB is open-source software and is freely available under its open-source license.

Source

https://github.com/vitrivr/cottontaildb