yggy

What is yggy?

yggy builds and maintains knowledge graphs from cluttered documents and systems without forcing you to design a schema or run a hand-built pipeline. Every fact keeps its source and time; duplicates can merge with undo; unclear matches go to human review. One graph powers the product UI, REST API, CLI, SDKs, and MCP for coding agents. Hosting options include EU cloud, private cloud, or on-prem. It targets developers and domain experts who want agent-ready company knowledge with evidence, not another black-box chat-over-docs tool.

How to use yggy?

1. Create a yggy project and connect sources such as docs, drives, or system exports.. Let extraction and entity matching populate the graph; review unclear merges in the queue.. Explore entities and source-backed facts in the app.. Connect Claude Code, Cursor, or Codex via MCP (or use the API/CLI).. Ask questions that need precise entities and citations, then refine ontology as needed.

Core Features

• Auto knowledge graph — Extract and match entities without designing a schema first.• Source and time on every fact — Answers cite documents and validity windows.• Reversible merges — Undo bad entity resolution from a Resolved view.• Human review queue — Route ambiguous matches to domain experts.• MCP, API, CLI, SDKs — Same graph for agents and application code.• Multi-tenant security — Isolated orgs/projects, SSO, scoped tokens, audit log.• EU / private / on-prem options — Keep knowledge where compliance requires.

Use Cases

• Ground coding agents on internal parts, contracts, and policies with citations.• Merge ERP and QMS identifiers for the same physical component.• Replace chunk-only RAG when questions need relationship constraints.• Let domain experts approve entity matches before agents rely on them.• Query supplier, leave, and policy knowledge together in one evidence trail.