Mnemosyne-OS/Mnemosyne-Neural-OS

Overview

A self-hosted memory layer for AI models. It indexes files and conversations into searchable vaults stored on your hardware, and connected models can query that memory. You control what gets added, moved, or removed. Runs on Windows, macOS, and Linux.

Best for

Best for
Developers who want a self-hosted, model-agnostic memory layer for AI tools.

Use cases

• Index local files and chat logs for model retrieval
• Give connected AI models persistent, searchable context
• Manage and curate memory vaults on your own hardware

How to use

Install

npx -y @mnemosyne_os/mcp

Tested with

Cursor, ChatGPT

Notes

A self-hosted memory layer for AI models. It indexes files and conversations into searchable vaults stored on your hardware, and connected models can query that memory. You control what gets added, moved, or removed. Runs on Windows, macOS, and Linux.

14 stars on GitHub. Last updated 2026-09-16.

Use cases

• Index local files and chat logs for model retrieval
• Give connected AI models persistent, searchable context
• Manage and curate memory vaults on your own hardware

Pros

• Data stays on your hardware
• Model-agnostic: works with any model you connect
• Cross-platform support

Cons

• Small project with limited community (14 stars)
• Requires manual curation of what enters and leaves
• No hosted option; you run it yourself

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.