gingugu/gingugu

What it does

gingugu is a persistent memory system designed specifically for AI coding assistants. It provides local, cloud-free storage using SQLite, allowing AI tools to maintain context and knowledge across sessions. The system manages typed memories with confidence lifecycles (verified/inferred/stale/deprecated), supports namespaces, and builds knowledge graphs for intelligent information retrieval.

How to set up

gingugu is deployed as an MCP (Model Context Protocol) server. Installation involves cloning the repository from GitHub and configuring it as an MCP server with your preferred AI assistant (Claude, ChatGPT, or others supporting MCP). The system uses local SQLite for storage and integrates with OS keychains for secure credential management, requiring no external cloud services or API keys for the core functionality.

Key features

• 16 MCP tools including store, recall, search, relate, consolidate, and export functions
• Local SQLite database with no cloud dependency
• Typed memory system with confidence lifecycle tracking
• Namespace support for organized information storage
• Knowledge graph capabilities for relationship mapping
• Hybrid search functionality for intelligent retrieval
• Integrated credential vault using OS keychain
• Cross-platform support (Windows, macOS, Linux)