Remembrances

About

Remembrances is a Go-based MCP (Multi-Agent Communication Protocol) server designed to equip AI agents with robust long-term memory. It supports multiple memory layers, including key-value stores, vector/RAG for semantic search, and graph databases, leveraging SurrealDB for data persistence. Additionally, it enables comprehensive knowledge base management via Markdown files and offers flexible embedding generation with local GGUF models, Ollama, or OpenAI API, alongside a powerful code indexing system.

Key Features

• Manages knowledge bases through Markdown files for AI agent access
• Includes a Code Indexing System utilizing Tree-sitter for multi-language AST parsing, semantic search, and code manipulation
• Offers an intelligent help system for on-demand tool documentation, reducing initial context token consumption
• Generates embeddings with local GGUF models (privacy-first, GPU accelerated), Ollama, or OpenAI API
• Equips AI agents with multi-layered long-term memory (key-value, vector/RAG, graph) using SurrealDB
• 0 GitHub stars

Use Cases

• Empowering AI agents with persistent, queryable memory for complex tasks and interactions
• Indexing and semantically searching codebases to assist AI agents in development and understanding
• Providing AI agents with access to managed knowledge bases for informed decision-making