Tutorial: Building with CleaveDB

Introduction to CleaveDB

Welcome to the CleaveDB tutorial. CleaveDB is a next-generation, high-performance hybrid database that brings together the best aspects of document stores, graph databases, and vector search engines into a single, unified architecture. Designed from the ground up for the AI era, it eliminates the need to stitch together multiple disparate systems.

Unlike traditional relational databases that require rigid schemas and expensive JOIN operations, CleaveDB embraces a flexible document model. Data is stored in isolated tenant namespaces, ensuring that multi-tenant applications remain secure and performant. Relationships between data points are handled through native graph bonds, allowing you to traverse complex connections instantly without the overhead of junction tables.

Furthermore, CleaveDB is built with native AI capabilities. It features an embedded ONNX Transformer model that automatically generates high-dimensional vector embeddings for your data. This allows you to perform deep semantic searches, understanding the meaning behind the text rather than just matching exact keywords, all within the same query execution pipeline.

In this tutorial, we will explore how to model data, establish connections, and query information efficiently using CleaveQL, our conversational query language designed for developer ergonomics and extreme performance.