aayoawoyemi/Ori-Mnemos

Local-first persistent agentic memory powered by Recursive Memory Harness (RMH). Open source must win.

What it solves

Ori Mnemos provides a persistent, local-first memory infrastructure for AI agents. It eliminates the need for cloud-dependent vector databases or vendor lock-in by allowing agents to store and retrieve information using plain markdown files on disk, ensuring the user owns all their data.

How it works

Ori implements the Recursive Memory Harness (RMH) framework, treating memory as a navigable knowledge graph rather than a flat search index. It uses markdown files as nodes and wiki-links as edges. The system employs mathematical models of human cognition, including ACT-R activation decay (to forget irrelevant information) and Hebbian learning (to strengthen connections between notes retrieved together).

Retrieval is handled through a combination of BM25, embeddings, and PageRank fusion. For complex queries, the system uses "navigated recursion," where it decomposes a question into sub-questions and traverses the graph to find answers. It is accessible via a CLI, a programmatic library, and an MCP (Model Context Protocol) server for direct agent integration.

Who it’s for

It is designed for developers building AI agents that require long-term, cross-session memory that is private, local, and capable of evolving based on usage patterns.

Highlights

  • Local-First Architecture: Uses markdown files and SQLite, requiring no API keys or cloud infrastructure for retrieval.
  • Cognitive Forgetting: Implements ACT-R base-level learning equations to naturally decay and forget unimportant information.
  • Graph-Based Retrieval: Uses Personalized PageRank and spreading activation to navigate memory via wiki-links.
  • Recursive Exploration: Decomposes unresolved queries into sub-questions to steer the traversal of the knowledge graph.
  • Agent Integration: Provides bridges for Claude Code, Hermes, OpenCode, and any MCP-compatible client.

Related

  • Project
    doobidoo/mcp-memory-service

    A self‑hosted memory backend for LLM agents that provides a REST API, CLI, OAuth, and a web dashboard. It stores facts, decisions, and causal knowledge graphs locally (SQLite + ONNX embeddings) or in hybrid/cloud mode, offering fast (≈5 ms) semantic search, tag‑scoped retrieval, and optional LLM‑generated summaries. Works with LangGraph, CrewAI, AutoGen, Claude Desktop, ChatGPT, and many other tools.

  • Project
    agentscope-ai/ReMe

    A local-first, self-evolving personal knowledge base that stores AI agent memory as editable Markdown files, enabling long-term knowledge retention and cross-agent sharing.

  • Project
    MemoriLabs/Memori

    A persistent, structured memory layer for AI agents that captures conversations and and agent execution traces to eliminate session amnesia and reduce token costs.

  • Project
    ThinkfleetAI/memmesh

    A persistent, local-first memory store for AI agents that provides durable, typed, and searchable memory using a single Rust binary without requiring a hosted vector database.

  • Project
    MemTensor/MemOS

    MemOS 2.0 is a Memory Operating System for LLM‑based AI agents. It offers a unified API to store, retrieve, edit, and delete multi‑modal memories (text, images, tool traces) organized as a graph. The system can run as a hosted cloud service (API‑key access) or be self‑hosted via Docker (Neo4j + Qdrant) or a local SQLite plugin. Ready‑made plugins integrate MemOS with OpenClaw, Hermes, and DeepSeek Harness. Benchmarks show it outperforms existing commercial memory products. Licensed under Apache 2.0.