# Dataojitori/nocturne_memory

A lightweight, rollbackable, and visual Long-Term Memory Server for MCP Agents. Say goodbye to Vector RAG and amnesia. Empower your AI with persistent, graph-like structured memory across any model, session, or tool. Drop-in replacement for OpenClaw.

Repository: https://github.com/Dataojitori/nocturne_memory
Canonical: https://ross.abutalabs.com/products/nocturne_memory
Homepage: https://misaligned.top/memory
Language: Python
License: MIT
License Family: permissive
Topics: agentic-ai, ai-memory, artificial-intelligence, long-term-memory, mcp, mcp-server, python, sqlite, ai-identity, claude, claude-code, digital-soul, llm, rag, second-brain, postgresql, gemini-cli
Last push: 2026-08-09T18:02:36+00:00

## Health v2 (maintenance only)
Score: 80/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 96, release rhythm 96, longevity 17
- inputs: {"age_days": 251, "days_push": 24, "days_rel": 24, "gap_med": 4, "n_releases_24m": 26}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1333, forks 166 (observed 2026-08-28T04:04:24.483533+00:00)

## What it is
Nocturne Memory is a lightweight, rollbackable long-term memory server for MCP agents, storing persistent, graph-like structured memories in SQLite or PostgreSQL. It acts as a drop-in replacement for OpenClaw, letting AI assistants recall identity and context across sessions, models, and tools via the MCP protocol.

## Use cases
- give my AI agent persistent memory across chat sessions
- remember user preferences and identity across different LLM models
- replace vector RAG with structured long-term memory for agents
- add a memory MCP server to Claude Code or Cursor
- rollback an AI assistant's memory to a previous state
- build a second-brain memory store for my coding assistant

## When to choose
- you use MCP-compatible clients (Claude Code, Cursor, Gemini CLI, etc.) and need cross-session memory
- you want structured, graph-like memory instead of embedding-based vector RAG
- you need a lightweight self-hosted memory backend with SQLite or PostgreSQL
- you want memory rollback and visual inspection of agent memory

## When to avoid
- you need large-scale semantic search over massive document corpora, where vector databases fit better
- you require a managed cloud service rather than self-hosting
- your stack has no MCP support and you cannot add it

## Facets
- artifact type: service
- maturity: active
- function: database, search-engine, mcp, rag, chatbot
- domain: artificial-intelligence, large-language-models, databases, developer-tools
- platform: python, self-hosted, cross-platform, cli
- tags: mcp-server, long-term-memory, ai-memory, sqlite, postgresql, second-brain, agent-memory, persistent-memory, graph-memory, claude, ai-agents

## Member repositories
- Dataojitori/nocturne_memory (main) score 80

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:24.483533+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:44:39.502532+00:00, confidence not recorded.
  - readme: https://github.com/Dataojitori/nocturne_memory (fetched 2026-08-28T04:04:24.483533+00:00, sha 53c890abe683)
  - homepage: https://misaligned.top/memory (fetched 2026-08-29T12:03:58.960293+00:00, sha c0f66347d255)
- Data as of 2026-08-30T08:39:29.467469+00:00.