Ori-Mnemos

概述

Ori Mnemos实现了一个复杂的基于知识图谱的记忆系统，使用数学模型模拟人类认知。它包含激活衰减、沿wiki链接的激活传播、赫布共现学习和检索模式的强化学习。该系统本地运行，零云依赖，数据存储在markdown文件中并使用SQLite索引。其递归记忆框架(RMH)将检索视为导航而非搜索，通过图遍历和子问题分解实现多跳推理。

试试问 AI

装完之后，这里有 5 个你可以让 AI 做的事：

• add_note

  Add a new note to the memory vault

  试试:Add a note about my meeting with John today试试:Create a new note about the Q3 sales strategy
• search_memory

  Search through the memory vault for relevant information

  试试:Find information about the marketing campaign试试:Search for notes related to product development
• update_note

  Update an existing note in the memory vault

  试试:Update the note with my meeting notes to add action items试试:Revise the project timeline note
• explore_topic

  Perform recursive exploration of a topic using sub-question decomposition

  试试:Explore the topic of AI ethics in our industry试试:Deep dive into competitor analysis
• get_graph_stats

  Get statistics about the knowledge graph structure

  试试:Show me information about my knowledge graph试试:Get metrics on wiki-link connections
• promote_note

  Classify and promote a note from inbox to appropriate memory space

  试试:Promote this idea to my knowledge base试试:Classify this note as a decision
• get_identity

  Retrieve the agent's stored identity information

  试试:What is my agent's identity?试试:Show me my stored goals and methodology
• list_notes

  List notes in a specific memory space with filtering options

  试试:Show me all my notes in the knowledge space试试:List recent operations notes
• get_note

  Retrieve a specific note by ID or title

  试试:Get my note about the quarterly review试试:Show me the note with ID 12345
• delete_note

  Delete a specific note from the memory vault

  试试:Remove the outdated meeting notes试试:Delete the draft that's no longer needed
• create_wiki_link

  Create a wiki-link connection between two notes

  试试:Link my product notes to the strategy document试试:Connect this research to the relevant project
• prune_memory

  Analyze and archive notes based on activation topology

  试试:Clean up outdated notes from my memory试试:Prune low-value notes to optimize storage

可对比工具

安装

npm install -g ori-memory
ori init my-agent
cd my-agent

{
  "mcpServers": {
    "ori": {
      "command": "ori",
      "args": ["serve", "--mcp", "--vault", "/path/to/brain"],
      "env": { "ORI_VAULT": "/path/to/brain" }
    }
  }
}

FAQ

Ori与其他记忆系统相比如何？

Ori在基准测试中显著优于Mem0等系统，召回率高3.1倍，F1分数高2.1倍，速度快9.5倍—all仅使用本地markdown文件和SQLite，无需云服务。

Ori支持哪些存储后端？

Ori使用磁盘上的markdown文件进行存储，SQLite用于索引和向量嵌入。它没有数据库锁定限制，可与任何文件系统配合使用。

Hacker News 讨论

开发者社区最近的相关讨论。

• Ori Mnemos – an open-source persistent memory for AI agents▲ 2💬 1
  帖子 by starro____ · 2026-03-03
• 3AM Coding:cracking persistent open-source memory for agents▲ 1
  帖子 by starro____ · 2026-03-09
• Open Source, Decentralized Memory Layer for AI Agents▲ 1
  帖子 by starro____ · 2026-03-04

Ori-Mnemos 对比