Happy slice of mcp pizzaMCP.pizza

membase-mcp

MCP.Pizza Chef: unibaseio

Membase-MCP is a decentralized memory server that enables AI agents to securely store and retrieve conversation histories and interaction records via the Unibase memory layer. It ensures persistent, verifiable, and personalized agent memory across sessions, supporting agent continuity and traceability. The server integrates with the Membase protocol to interact with the Unibase DA network, providing a lightweight and secure memory solution for AI workflows.

Use This MCP server To

Store AI agent conversation history persistently and securely Retrieve past interaction records for agent continuity Switch between multiple conversation contexts dynamically Save and manage memories in decentralized storage Enable personalized AI agent responses based on stored memories Provide verifiable and tamper-proof memory storage for agents Integrate decentralized memory into AI workflows Support multi-session AI agent memory synchronization

README

membase mcp server

Description

Membase is the first decentralized memory layer for AI agents, powered by Unibase. It provides secure, persistent storage for conversation history, interaction records, and knowledge — ensuring agent continuity, personalization, and traceability.

The Membase-MCP Server enables seamless integration with the Membase protocol, allowing agents to upload and retrieve memory from the Unibase DA network for decentralized, verifiable storage.

Functions

Messages or memoiries can be visit at: https://testnet.hub.membase.io/

  • get_conversation_id: Get the current conversation id.
  • switch_conversation: Switch to a different conversation.
  • save_message: Save a message/memory into the current conversation.
  • get_messages: Get the last n messages from the current conversation.

Installation

git clone https://github.com/unibaseio/membase-mcp.git
cd membase-mcp
uv run src/membase_mcp/server.py

Environment variables

  • MEMBASE_ACCOUNT: your account to upload
  • MEMBASE_CONVERSATION_ID: your conversation id, should be unique, will preload its history
  • MEMBASE_ID: your instance id

Configuration on Claude/Windsurf/Cursor/Cline

{
  "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"
      }
    }
  }
}

Usage

call functions in llm chat

  • get conversation id and switch conversation

get conversation id and switch conversation

  • save message and get messages

save message and get messages

membase-mcp FAQ

How do I install the Membase-MCP server?
Clone the GitHub repo and run the server script using 'uv run src/membase_mcp/server.py'.
What is the Unibase memory layer?
Unibase is a decentralized data network that provides secure, verifiable storage for AI agent memories.
How does Membase-MCP ensure memory security?
It uses decentralized storage on the Unibase DA network, making memories tamper-proof and verifiable.
Can I switch between different conversations?
Yes, the server supports switching conversation contexts via the 'switch_conversation' function.
How do I retrieve past messages?
Use the 'get_messages' function to fetch the last n messages from the current conversation.
Is Membase-MCP compatible with multiple AI models?
Yes, it is model-agnostic and can be integrated with various LLM providers like OpenAI, Claude, and Gemini.
What environment variables are required?
Specific environment variables are needed for configuration; refer to the GitHub repo for details.
Can Membase-MCP handle multiple concurrent conversations?
Yes, it supports managing multiple conversation IDs and switching between them seamlessly.