Membase

Provides decentralized memory storage and retrieval for AI agents, enabling persistent conversation history and memory management via the Membase MCP interface.

python

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": {
    "unibaseio-membase-mcp": {
      "command": "uv",
      "args": [
        "--directory",
        "path/to/membase-mcp",
        "run",
        "src/membase_mcp/server.py"
      ],
      "env": {
        "MEMBASE_ID": "your sub account, any string",
        "MEMBASE_ACCOUNT": "your account, 0x...",
        "MEMBASE_CONVERSATION_ID": "your conversation id, should be unique"
      }
    }
  }
}

Membase MCP Server provides a secure, decentralized memory layer for AI agents, enabling persistent storage of conversation history, interaction records, and knowledge. It integrates with the Membase protocol to upload and retrieve memories from the Unibase network, fostering agent continuity, personalization, and traceability.

How to use

You interact with Membase MCP Server through an MCP client to manage conversations and memories. Use the available functions to keep track of a single conversation, switch between conversations, save new memories, and retrieve recent memories for context.

How to install

Prerequisites: you need the runtime used by the server (uv) and a working environment where you can run Python files. You will also need a Membase account, a conversation identifier, and your instance identifier set as environment variables.

# Step 1: Clone the MCP server repository
git clone https://github.com/unibaseio/membase-mcp.git

# Step 2: Change into the project directory
cd membase-mcp

# Step 3: Run the MCP server
uv run src/membase_mcp/server.py

Configuration and environment

Configure the MCP client to connect to this server using the following example. It shows the standard MCP JSON snippet you would place in your client configuration to start the server locally via the uv runtime.

{
  "mcpServers": {
    "membase": {
      "command": "uv",
      "args": [
        "--directory",
        "path/to/membase-mcp",
        "run", 
        "src/membase_mcp/server.py"
      ],
      "env": {
        "MEMBASE_ACCOUNT": "your account, 0x...",
        "MEMBASE_CONVERSATION_ID": "your conversation id, should be unique",
        "MEMBASE_ID": "your sub account, any string"
      }
    }
  }
}

Use the available functions

These are the core actions you can perform through the MCP client to manage memory and conversations.

get_conversation_id  - Retrieve the current conversation ID
switch_conversation  - Switch to a different conversation
save_message       - Save a message or memory into the current conversation
get_messages       - Retrieve the last n messages from the current conversation

Available tools

get_conversation_id

Retrieve the current conversation ID from the server.

switch_conversation

Switch to a different conversation to continue memory management in that thread.

save_message

Save a message or memory into the current conversation to persist history.

get_messages

Get the last n messages from the current conversation to provide context for following interactions.

Built by

VeilStrat

AI signals for GTM teams