MCP server with persistent memory and FTS5 search for Claude

An open-source MCP server for Claude, featuring persistent memory, FTS5 search on conversation history, and a web UI.

Claudex: MCP Server for Claude with Persistent Memory and FTS5 Search

Claudex is an open-source MCP server designed to enhance the capabilities of Claude, an AI assistant. This tool provides developers with a robust platform for building more sophisticated AI applications by integrating persistent memory and advanced search functionalities directly into the Claude interaction loop. By leveraging FTS5 for efficient searching of conversation history, Claudex enables richer context management and more intelligent responses from your Claude instances.

What Claudex Does

Claudex acts as a middleware, connecting your Claude instances to a persistent storage layer and providing a powerful search engine for conversational data. It allows Claude to "remember" past interactions beyond the typical context window, enabling more coherent and contextually aware conversations over extended periods. The integrated FTS5 search engine allows for rapid retrieval of relevant information from this stored history, making it possible to build applications that can recall specific details or themes from previous exchanges.

Key Features

Persistent Memory: Stores conversation history and relevant data, allowing Claude to maintain context across sessions.
FTS5 Search: Implements Full-Text Search 5 for fast and efficient querying of conversation logs.
Web UI: Provides a user-friendly interface for managing and interacting with the server and Claude.
Open-Source: Developed and maintained on GitHub, offering transparency and extensibility for developers.
MCP Server: Designed to integrate seamlessly with existing MCP-based AI development workflows.

Who Claudex is For

Claudex is specifically built for AI developers and engineers working with Claude. It is ideal for those looking to:

Build AI agents with long-term memory.
Develop conversational applications requiring recall of past interactions.
Integrate Claude into systems that need to reference historical data.
Experiment with advanced context management techniques for large language models.
Contribute to or customize an open-source AI infrastructure.