# Mnemodb

> Mnemodb is an open-source MCP server for knowledge and memory. Agent memory kept as plain Markdown files in your repo, with provenance and lifecycle. It runs locally on your machine and is started by your AI client.

- **Kind:** MCP server
- **Pricing:** Open source
- **Platforms:** CLI, macOS, Linux, Windows
- **Repository:** https://github.com/mnemodb/mnemodb
- **MCP Registry name:** dev.mnemodb/mnemodb
- **Categories:** [knowledge and memory](https://ai-powered.com/categories/knowledge-and-memory)
- **Listing updated:** 2026-11-18

## About

MnemoDB defines a .mem.md format where each memory is a Markdown heading with a one-line metadata span recording an id, source, confidence and what it supersedes. Because it is plain text, it renders anywhere and diffs in git. Existing CLAUDE.md or AGENTS.md files already count as valid untyped files.

Provenance is the core idea. Every entry is marked as coming from the user, the agent or a tool. Tool-sourced entries come back flagged untrusted, cannot be pinned into every session and cannot supersede or retire a trusted entry. The README is candid that the agent applies the label itself, so it catches honest mistakes rather than a determined injection.

Memories can carry TTLs and supersession, and a compact pass archives expired or replaced ones. The README says it is early (v0.1) and may be overkill if a short hand-kept CLAUDE.md already does the job.

## Features

- **Plain Markdown store:** A .memory/ folder of .mem.md files you can read, edit and diff.
- **Provenance labels:** Entries record user, agent or tool origin; tool entries are untrusted on recall.
- **Lifecycle rules:** TTLs, supersession and a compact pass that archives expired entries.
- **Auditable forgetting:** memory_forget is trust-gated and never erases a higher-trust memory.
- **Store health checks:** The mnemo doctor command lints for staleness, contradictions and damage.
- **Cross-tool portability:** Works with Claude Code, Cursor and other MCP clients.

## Good for

- Keep decisions and preferences across Claude Code sessions
- Audit what an agent learned with git diff
- Trace everything saved from tool output after a suspected injection
- Share one memory store between Claude Code and Cursor

## Install

### Claude Code (Terminal)

```sh
claude mcp add mnemodb -- npx -y @mnemodb/mcp
```

### Claude Desktop (claude_desktop_config.json)

```json
{
  "mcpServers": {
    "mnemodb": {
      "command": "npx",
      "args": [
        "-y",
        "@mnemodb/mcp"
      ]
    }
  }
}
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "mnemodb": {
      "command": "npx",
      "args": [
        "-y",
        "@mnemodb/mcp"
      ]
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "mnemodb": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "@mnemodb/mcp"
      ]
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "mnemodb": {
      "command": "npx",
      "args": [
        "-y",
        "@mnemodb/mcp"
      ]
    }
  }
}
```

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "mnemodb": {
      "command": "npx",
      "args": [
        "-y",
        "@mnemodb/mcp"
      ]
    }
  }
}
```

## Tools

- `memory_recall`: Recall stored memories.
- `memory_list`: List memories.
- `memory_show`: Show one memory entry.
- `memory_history`: Show an entry's supersession lineage.
- `memory_stats`: Report on the store itself.
- `memory_remember`: Save a memory.
- `memory_forget`: Forget a memory auditably, gated by trust.
- `memory_pin`: Control what loads into context.
- `memory_review`: Review memories.
- `memory_compact`: Archive expired or replaced memories.
- `memory_boot`: Boot-time memory loading.

## FAQ

### How do I install it?

Run claude mcp add mnemodb -- npx -y @mnemodb/mcp@latest, or use the Claude Code plugin.

### What is the licence?

Apache-2.0 for the code and CC BY 4.0 for the spec.

### Where is memory stored?

In a .memory/ directory chosen by MNEMO_STORE, or Claude Code's project directory, or the working directory.

### How many tools does it expose?

Eleven memory tools, from memory_recall to memory_boot.

### Is Mnemodb free?

Yes. Mnemodb is open source, so you can use and self-host it for free.

### How do I add Mnemodb to Claude Code?

Run this command in your terminal: claude mcp add mnemodb -- npx -y @mnemodb/mcp

### What are the best alternatives to Mnemodb?

The closest alternatives to Mnemodb are Codebase Memory, XERJ, Claude Code Ultimate Guide, Nautilus Compass Memory and NeuroStack.

## Where this comes from

- Description: written from the project's README.
- Install steps: from the official MCP Registry entry.

## Alternatives

- [Codebase Memory](https://ai-powered.com/mcp/codebase-memory): Codebase Memory is an open-source MCP server for knowledge and memory.
- [XERJ](https://ai-powered.com/mcp/xerj): XERJ is an open-source MCP server for knowledge and memory.
- [Claude Code Ultimate Guide](https://ai-powered.com/mcp/claude-code-ultimate-guide): Claude Code Ultimate Guide is an open-source MCP server for ai models and llm tooling.
- [Nautilus Compass Memory](https://ai-powered.com/mcp/nautilus-compass-memory): Nautilus Compass Memory is an open-source MCP server for knowledge and memory.
- [NeuroStack](https://ai-powered.com/mcp/neurostack): NeuroStack is a freemium MCP server for knowledge and memory.
- [note1](https://ai-powered.com/mcp/note1): note1 is an MCP server for productivity.
- [Digitalnz](https://ai-powered.com/mcp/digitalnz): Digitalnz is a freemium MCP server for knowledge and memory.
- [Notion_connect](https://ai-powered.com/mcp/notion-connect): Notion_connect is a freemium MCP server for knowledge and memory.

---
Source: https://ai-powered.com/mcp/mnemodb
