MCPFast / Tools / Local MCP Memory Server without API or LLM
A local MCP memory server running on Bun, using SQLite, requiring no API keys, LLMs, or embedding providers.
View on GitHub→This tool provides a self-contained, local solution for managing MCP (Memory, Context, and Persistence) data. Designed for developers who need direct control over their data and infrastructure, it eliminates reliance on external APIs, LLMs, or embedding services. By running on Bun and utilizing SQLite, it offers a lightweight and efficient way to store and retrieve contextual information for your AI applications. This approach is ideal for scenarios where data privacy, cost optimization, or offline functionality are critical requirements.
The Local MCP Memory Server acts as a dedicated backend for storing and retrieving structured and unstructured data relevant to your AI agents and applications. It handles the persistence of information, allowing your AI to maintain context across interactions without needing to re-fetch or re-process data from external sources. This local server is built to be straightforward, focusing on the core functionality of memory management.
This tool is targeted at AI developers, researchers, and hobbyists who are building applications that require persistent memory and context management. It is particularly useful for: