Memory-Plus

Provides a local RAG memory store for MCP agents with recording, retrieval, updating, and visualization of memories.
  • python

50

GitHub Stars

python

Language

6 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs
{
  "mcpServers": {
    "yuchen20-memory-plus": {
      "command": "uvx",
      "args": [
        "-q",
        "memory-plus@latest"
      ],
      "env": {
        "GOOGLE_API_KEY": "YOUR_API_KEY"
      }
    }
  }
}

Memory-Plus is a lightweight local memory store for MCP agents. It lets your agent record, search, update, and visualize persistent memories across conversations, helping maintain context and build richer interactions over time.

How to use

You will run Memory-Plus as an MCP server locally and connect it from your MCP client. Start Memory-Plus, then add it as an MCP server in your client configuration. Once connected, you can: record memories from conversations, retrieve memories by keywords or topics, fetch recent memories, update existing entries, visualize memory relationships, import documents into memory, delete entries, and leverage memory versioning to review past states.

How to install

Prerequisites you need to prepare before starting Memory-Plus:

  • A valid Google API key to access the Gemini Embedding API; set it as GOOGLE_API_KEY in your environment
  • Python and the UV runtime installed on your system
Install the UV runtime if you have not yet:
- Run: `pip install uv

Optional setup for rapid local testing with MCP Inspector:

  • Clone Memory-Plus locally
  • Use MCP Inspector to run the server in a fast MCP workflow as described in the local testing workflow

Configure and start Memory-Plus through the MCP shell or an IDE by using the stdio configuration shown below. The recommended runtime command to start Memory-Plus in MCP environments is:uvx -q memory-plus@latest

Configuration and usage notes

Connection method: stdio (local server) using UVX. The command line to start Memory-Plus is:

uvx -q memory-plus@latest

This starts Memory-Plus and makes it available to your MCP client as an MCP server.

Environment variables you should provide (example): GOOGLE_API_KEY=YOUR_API_KEY Set this in your environment before starting the server so memory embedding and related features work properly.

Server configuration (MCP JSON snippet)

{
  "mcpServers": {
    "memory_plus": {
      "type": "stdio",
      "name": "memory_plus",
      "command": "uvx",
      "args": ["-q", "memory-plus@latest"],
      "env": [
        {"name": "GOOGLE_API_KEY", "value": "YOUR_API_KEY"}
      ]
    }
  }
}

Security and maintenance

Keep your GOOGLE_API_KEY secure and do not expose it in public environments. Regularly update Memory-Plus to include the latest features like memory versioning and improved memory graph visualizations.

Available tools

Record Memories

Save user data, ideas, and important context as persistent memories for future sessions.

Retrieve Memories

Search memories by keywords or topics across past entries to retrieve relevant context.

Recent Memories

Fetch the last N memory entries quickly for quick context recall.

Update Memories

Append or modify existing memories to reflect new information.

Visualize Memories

Display interactive memory graphs showing relationships between entries.

File Import

Ingest documents directly into memory to enrich context.

Delete Memories

Remove unwanted entries from memory.

Memory for Memories

Control when and how past memories are recalled using resources.

Memory Versioning

Keep historical versions of memory entries to preserve full history.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational