mnemo-cortex
Active·★ 158·MIT·Updated 2026-10-21
Mnemo Cortex provides AI agents with persistent, local, and cross-agent memory, enabling deep recall, overnight synthesis, and structured fact management without cloud dependencies. It functions as a cognitive coprocessor, enhancing agent intelligence by automatically capturing and consolidating conversational context across multiple tools and sessions.
#Active Memory#AI Agents#Cognitive Coprocessor#Local-first#Cross-Agent Memory#Semantic Search#LLM Memory#Data Persistence
01
Features
01Active Memory: Automatically captures, classifies, and consolidates agent interactions, learning what works over time.
02Cross-Agent Dreaming: Synthesizes recent memories across all agents overnight, creating shared context for collective intelligence.
03Structured Facts: Provides a key-value store with confidence tracking for sub-millisecond exact lookup of entities and attributes, enhancing knowledge accuracy.
04Local-first & No Cloud: Operates entirely on your local machine, guaranteeing privacy and independence from cloud services.
05Document Discovery (The Librarian): Indexes all workspace files (including content snippets) to enable agents to quickly find documents based on fuzzy descriptions.
02
Compatibility
Python
Runtime
Verified via docs
Ollama
LLM Provider
Verified via docs
Node.js
Bridge Runtime
Verified via docs
Linux
OS
Verified via docs
macOS
OS
Verified via docs
Windows
OS
Verified via docs
03
Quick start
1
$ git clone https://github.com/GuyMannDude/mnemo-cortex.git
2
$ cd mnemo-cortex
3
$ python -m venv .venv
4
$ source .venv/bin/activate
5
$ pip install -e .
04
Use cases
↳Providing AI agents with persistent, long-term memory that survives sessions and tool switches, preventing amnesia.
↳Facilitating seamless collaboration among multiple AI agents by sharing and synthesizing learned context overnight, making collective knowledge available.
↳Creating a private, self-hosted knowledge base for AI agents, ensuring data privacy and eliminating reliance on external cloud services.
↳Storing and retrieving structured factual information with confidence tracking for precise lookups, complementing semantic search for critical data.
↳Safely ingesting and managing developer behavioral claims into an agent's context, including review queues and content detectors for robust privacy and security.
05
Alternatives
mcp-local-memory★ 60
A lightweight, powerful local memory server for AI agents supporting text, entities, and relations. Enables persistent codebase understanding and user preference management.
brain.md★ 34
Local-first markdown notes with a first-class MCP server. 16 tools + 2 resources for Claude Code, Claude Desktop, Cursor and any MCP agent. Per-folder permissions, semantic RAG via LanceDB, single Bun binary. AGPL.
infimium-agent★ 25
Infimium helps agents find the right code, docs, and dependencies — before they edit. Local-first, open-source, MCP-compatible.
Compartment★ 584
Encrypted, fully offline agentic memory. One click install, GUI w/ memory map, all OS and agents. Logically and mathematically sound storage and retrieval.
mneme★ 167
🧠 The memory that dreams — cross-session memory for DeepSeek Harness. Offline & private, auto-consolidates in its sleep (autoDream), visualized in a memory panel.
Comments
Log in to leave a comment
No comments yet. Be the first!