Skip to content

Acknowledgments

The Madeinoz Knowledge System is built on the shoulders of giants. This page acknowledges the projects, communities, and individuals who made this system possible.

Core Dependencies

Graphiti by Zep

Graphiti is the foundational knowledge graph engine that powers this system. It provides:

  • Temporal Knowledge Graph: Bi-temporal tracking with created_at and last_updated timestamps
  • Entity Extraction: Automatic extraction of entities, relationships, and episodes from text
  • Vector Embeddings: Semantic search capabilities using OpenAI embeddings
  • Neo4j/FalkorDB Backend: Flexible graph database support

Graphiti is an open-source library by Zep for building long-term memory for AI agents.

Neo4j

Neo4j is the default graph database backend, providing:

  • Native graph storage and querying with Cypher
  • Vector similarity search for semantic retrieval
  • ACID transactions for reliable knowledge storage
  • Neo4j Browser for visual graph exploration

FalkorDB

FalkorDB is the alternative graph database backend, providing:

  • Redis-based graph storage with RediSearch
  • Lucene-based full-text and vector search
  • In-memory performance for fast queries
  • Built-in web UI at port 3000

Inspiration & Community

Personal AI Infrastructure (PAI)

PAI is the conceptual foundation for this knowledge system. The PAI project demonstrates how AI assistants can maintain persistent, queryable memory across conversations.

Specifically, we'd like to acknowledge:

  • PAI Discussion #527: Knowledge System Long-term Memory Strategy - This community discussion directly inspired the Memory Decay Scoring & Lifecycle Management feature (Feature 009). The concepts of importance classification, stability scoring, and automated memory lifecycle management emerged from these conversations.

The PAI community's approach to: - Conversational Knowledge Capture: Storing memories without manual organization - Automatic Entity Extraction: Using LLMs to identify people, organizations, and concepts - Relationship Mapping: Tracking how entities connect across episodes - Semantic Search: Finding information by meaning, not just keywords

...forms the core philosophy of this system.

Memory Systems Research

This system incorporates research and concepts from academic and industry sources:

Development Tools

Bun

Bun is the TypeScript runtime and package manager used for CLI tools and server orchestration.

MCP (Model Context Protocol)

MCP provides the standard protocol for communication between the knowledge system and AI assistants like Claude Code.

MkDocs Material

Documentation theme and build system for this documentation site.

Community & Contributors

This system exists to serve the PAI community and the broader AI assistant ecosystem. We extend our gratitude to:

  • Daniel Miessler for creating and maintaining PAI
  • The PAI community for feedback, discussions, and feature ideas
  • Zep for building and open-sourcing Graphiti
  • All contributors who report issues, submit PRs, and improve the system

License

This project is licensed under the MIT License.

The Madeinoz Knowledge System is a PAI Pack - a pluggable component for the Personal AI Infrastructure ecosystem.

Join the Community: