K-Memory Offers Persistent Memory for AI Agents
TL;DR. K-Memory 2.1 is a new pure Python memory engine designed for AI agents and LLMs. - It offers persistent storage, linking, summarizing, and knowledge export functions with zero dependencies. - The system includes features like recency boosting, auto-summaries, and native deduplication. - An optional plugin provides semantic search capabilities using local Ollama model integration.
- K-Memory v2.1 is a self-contained persistent memory engine for AI agents and LLMs.
- The system operates with zero external dependencies and is written entirely in Python.
- Key features include weighted recency, automatic domain summaries, and fact deduplication.
- It allows knowledge graph export in Markdown or Mermaid formats for visualization.
- An optional plugin enables semantic search for meaning, rather than keywords, using Ollama.
Sources
- K-Memory – Persistent Memory for AI — github.com