Skip to content
MCP server

MemVault - Local-First Agent Memory

Local-first shared memory layer that any MCP client can save to and recall from

Runs locallyStatus and history
Install
Pricing
Open source
Platforms
CLI, macOS, Linux
Runs
Locally
Version
0.7.2
Listing updated
2 days ago

About MemVault - Local-First Agent Memory

MemVault - Local-First Agent Memory is an open-source MCP server for knowledge and memory. It runs locally on your machine and is started by your AI client.

MemVault is a self-hosted router that keeps memories outside any one agent, so Claude Desktop, Cursor, Cline or other MCP clients share the same context. The README frames the idea as memory finding the agent: when a client connects, relevant stored items are injected as context, and entries marked MUST priority are injected as instructions the agent should follow.

A companion CLI saves, searches and lists memories, extracts structured items from free text, records task outcomes with lessons, deduplicates, decays and archives stale entries, and can generate agent instruction files from memory. Retrieval is hybrid keyword plus semantic when embeddings are enabled. It auto-detects a local Ollama for embeddings and LLM extraction, so it can run fully on your machine.

It is written in Rust and MIT licensed. Prebuilt Linux binaries need a recent glibc (2.39 or later); older systems are pointed to the Docker image or a source build. A local dashboard is started with memvault-cli dashboard.

Install

Run in a terminal

claude mcp add memvault -- docker run -i --rm ghcr.io/dreamor/memvault:v0.7.2

Add to claude_desktop_config.json

{
  "mcpServers": {
    "memvault": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "ghcr.io/dreamor/memvault:v0.7.2"
      ]
    }
  }
}
Install in CursorOpens Cursor, which asks you to confirm. Or add it by hand:

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "memvault": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "ghcr.io/dreamor/memvault:v0.7.2"
      ]
    }
  }
}
Install in VS CodeOpens VS Code, which asks you to confirm. Or add it by hand:

Add to .vscode/mcp.json

{
  "servers": {
    "memvault": {
      "type": "stdio",
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "ghcr.io/dreamor/memvault:v0.7.2"
      ]
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "memvault": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "ghcr.io/dreamor/memvault:v0.7.2"
      ]
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "memvault": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "ghcr.io/dreamor/memvault:v0.7.2"
      ]
    }
  }
}

What it does

Shared across agents
Any MCP client connecting to it sees the same memory.
Priority levels
MUST-level preferences are injected as instructions.
Hybrid retrieval
Keyword search plus semantic search when embeddings are on.
Local Ollama support
Embeddings and extraction can run on a local Ollama model.
Dedup and decay
Commands to dedupe, decay and archive stale memories.
Instruction file sync
A sync command generates agent instruction files from memory.

Good for

  • Keep coding preferences consistent across Claude Code and Cursor
  • Record lessons from failed deployments for later sessions
  • Extract memories from free text notes
  • Run agent memory entirely on a local machine

Pricing

Open source

MIT licensed, self-hosted.

MemVault - Local-First Agent Memory FAQ

Is it open source?

Yes, the README states MIT licence and self-hosted.

How is it installed?

Via an install script, Homebrew, cargo install, or the Docker image.

Does it need a cloud service?

No. Embeddings and extraction can use a local Ollama.

What are the Linux requirements for prebuilt binaries?

glibc 2.39 or later; older distributions should use Docker or build from source.

Is MemVault - Local-First Agent Memory free?

Yes. MemVault - Local-First Agent Memory is open source, so you can use and self-host it for free.

How do I add MemVault - Local-First Agent Memory to Claude Code?

Run this command in your terminal: claude mcp add memvault -- docker run -i --rm ghcr.io/dreamor/memvault:v0.7.2

What are the best alternatives to MemVault - Local-First Agent Memory?

The closest alternatives to MemVault - Local-First Agent Memory are Codebase Memory, XERJ, Claude Code Ultimate Guide, Nautilus Compass Memory and note1.

Alternatives to MemVault - Local-First Agent Memory

  1. Codebase Memory

    MCP server

    Local MCP server that builds a code knowledge graph for AI coding agents

    Runs locally
    • Open source
    • 48.7K GitHub stars
    • Knowledge and memory
  2. XERJ

    MCP server

    Local search engine for AI agents with cited answers over your code and docs

    Runs locally
    • Open source
    • 19.2K GitHub stars
    • Knowledge and memory
  3. MCP server for searching the Claude Code Ultimate Guide from any MCP client

    Runs locally
    • Open source
    • 6.5K GitHub stars
    • AI models and LLM tooling
  4. Local-first long-term memory layer for AI agents, served over MCP

    Not checked yet
    • 5 tools
    • Remote
    • Open source
    • 1.3K GitHub stars
    • Knowledge and memory
  5. note1

    MCP server

    MCP server giving AI assistants search and summaries across your note1 meetings

    • 4 tools
    • Remote
    • Free trial
    • 1 GitHub star
    • Productivity