Download
Synabun is an open-source AI memory management and augmentation system designed to provide persistent, semantic memory for AI agents and coding assistants, particularly those compatible with the MCP (Model Context Protocol) ecosystem. It functions as a local-first solution that stores and retrieves contextual knowledge across sessions using a built-in vector database powered by embeddings, eliminating the need for external APIs, cloud services, or Docker dependencies. The system integrates tightly with developer workflows by running alongside tools like Claude Code, enabling automatic memory capture, retrieval, and contextual augmentation through lifecycle hooks and commands. One of its defining characteristics is its Neural Interface, a browser-based 3D visualization that represents stored memories as nodes in an interactive graph, allowing users to explore relationships, edit entries, and manage knowledge visually.
Features
- Semantic vector search based on meaning rather than keywords
- Local-first architecture using SQLite and embedded models without external APIs
- Interactive 3D memory graph visualization interface
- Multi-project memory sharing with automatic context detection
- Automated lifecycle hooks for memory capture and retrieval
- File change detection using hashing to flag outdated knowledge
Project Samples
