seanmarkwei/agentram-mcp

What it does

AgentRAM-MCP is a Model Context Protocol (MCP) server that provides persistent memory capabilities for AI agents. It offers a lightweight, simple key-value HTTP API that allows agents to store and retrieve information without requiring complex vector databases or embedding systems. The tool enables multiple agents to share memory through namespaced storage and supports time-to-live (TTL) functionality for automatic data expiration.

How to set up

Installation is straightforward using npm: run npx y agentrammcp to get started. The tool is designed as an MCP server, making it compatible with Claude, ChatGPT, and other AI platforms that support the MCP standard. Once installed, you can begin storing and retrieving agent memory immediately through the HTTP API.

Key features

• Simple key-value storage for agent memory
• No vector database or embeddings required
• Persistent data storage across sessions
• Shared namespaces for multi-agent collaboration
• TTL (time-to-live) support for automatic data cleanup
• HTTP API for easy integration
• MCP server compatibility