yggy

Organizations store the same information across multiple systems, such as an ERP and a QMS, often under different identifiers, revisions and formats. A single component or fact can exist as two numbers with two revision states in two different systems, making it hard to know which version is current or valid. Teams and their AI agents need a way to ask questions and get a trustworthy answer backed by evidence, rather than manually reconciling documents, tables and systems. Today this reconciliation work is recurring and manual: extraction, modeling, entity matching and updates have to be redone as sources change.

yggy core automatically builds a knowledge graph from documents, tables and systems, turning them into entities and relationships without a custom pipeline project. It keeps the graph current: when a source changes, the graph updates, while earlier states remain visible as history. Every fact returned carries its source document and date, and domain experts step in to decide ambiguous or unclear cases rather than replacing existing systems. Developers and AI teams can connect to the graph via MCP from tools like Claude Code, Codex or Cursor, or use a CLI and API, so an agent can query entities and receive answers with citations. The company also offers yggy notes, a personal capture tool where the graph organizes notes automatically, and yggy quality, aimed at complaints and traceability work for quality teams.