Membase MCP

Membase MCP is an MCP server that provides a decentralized memory and context layer for AI agents powered by Unibase. It connects client applications to the Unibase Data Availability (DA) network, offering verifiable, persistent storage for conversation histories, interaction logs, and agent knowledge. Software developers and AI system builders use Membase MCP to maintain agent continuity, state tracking, and personalization across sessions without relying on centralized databases. By exposing tools to manage conversation identifiers and message logs, it enables large language models to store interaction records to the network and retrieve recent contextual histories on demand. The server interacts directly with decentralized storage layers, ensuring verifiable memory that can be checked independently. Agents can switch between multiple conversation threads, isolate contextual memory per session, and retrieve historical dialog segments to maintain consistent context during long-term interactions.

Category: AI Memory & Context

Tags: decentralized, knowledge-graph, memory, persistence

Visit Membase MCP

How to install and configure Membase MCP

1. Clone the repository locally using git: git clone https://github.com/unibaseio/membase-mcp.git. 2. Navigate into the cloned directory: cd membase-mcp. 3. Verify your environment can execute the server using uv run src/membase_mcp/server.py. 4. Open your MCP client configuration file (such as Claude Desktop, Cursor, Windsurf, or Cline). 5. Add the membase configuration under the mcpServers object: json { "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" } } } } 6. Replace the environment variables with your actual account address, conversation identifier, and instance ID, then restart your client.

What you can do with Membase MCP

• Persisting assistant chat history and memories across distinct sessions using the decentralized Unibase DA network. - Switching active conversation threads during an agent session to partition context for separate projects or tasks. - Retrieving the last n messages from a specific conversation thread to restore previous context accurately. - Verifying stored agent interaction records and knowledge logs publicly via the Membase testnet hub interface.

Key facts

• https://github.com/unibaseio/membase-mcp
• AI Memory & Context, Notes & Knowledge Management
• decentralized, knowledge-graph, memory, persistence

Part of MCP Servers

What is Membase MCP?

Membase MCP is a Model Context Protocol server that connects AI assistants to Membase, a decentralized memory network built on Unibase. It allows agents to read and write conversation histories and knowledge states to a decentralized storage layer for verifiable persistence.

How do I install Membase MCP?

Clone the repository from GitHub using git, then configure your MCP client with the uv command pointing to the repository path and src/membase_mcp/server.py. You must configure the MEMBASE_ACCOUNT, MEMBASE_CONVERSATION_ID, and MEMBASE_ID environment variables in your client configuration.

What tools does Membase MCP provide?

Membase MCP provides four primary functions: get_conversation_id to fetch the active ID, switch_conversation to change conversation contexts, save_message to record an interaction to decentralized storage, and get_messages to fetch the most recent messages from the current thread.

Which MCP clients work with Membase MCP?

Membase MCP works with standard MCP clients that support local command execution and environment variable injection, including Claude Desktop, Cursor, Windsurf, and Cline.