About this project

The googlecloudplatform/knowledge-catalog repository is a collection of tools, agents, and samples built around Google Cloud's Knowledge Catalog service (formerly known as Dataplex). Knowledge Catalog is an AI-powered data catalog and metadata management platform that constructs a dynamic knowledge graph spanning both structured and unstructured data assets. Its primary purpose is to provide semantics and business context to AI agents, enabling them to understand and reason about organizational data more effectively. The repository serves as a demonstration and reference library for Knowledge Catalog capabilities. It showcases how to build solutions for context management, data enrichment, and retrieval workflows. These solutions leverage the knowledge graph to surface relevant metadata, relationships, and business definitions that AI systems can use to improve accuracy and relevance. Key aspects of the repository include: - Tools and agents that interact with Knowledge Catalog to manage and query metadata - Sample implementations demonstrating context management patterns - Enrichment workflows that augment data with semantic information - Retrieval solutions that leverage the knowledge graph for data discovery The repository is designed to be opened and explored directly in Google Cloud Shell, making it easy for developers to experiment with the provided samples without local setup. It is licensed under Apache 2.0 and is explicitly noted as not being an official Google product, meaning it is a community-maintained resource rather than a first-party Google offering. The underlying Knowledge Catalog service addresses a common challenge in modern data environments: the difficulty of discovering, understanding, and trusting data across diverse sources. By maintaining a centralized knowledge graph with business context, it helps organizations bridge the gap between raw data and meaningful AI-driven insights. The repository's samples illustrate practical approaches to integrating this capability into existing data pipelines and AI agent architectures.