mcp -memory-file

This MCP server provides a lightweight memory store that reads and writes memories from a simple text file. Each line in the file represents one memory entry. The server exposes tooling like memory_add(memory: string), memory_search(query: string), memory_delete(query: string), memory_list(), and memory_update (an alias for memory_delete followed by memory_add). Claude or other MCP clients can use these tools to persist, retrieve, and prune memories during a chat session. For example, memory_add("Wes's daughter is 8") stores that memory, memory_search("Wes") returns matching memories, and memory_list() returns all stored memories. The model can also record tool usage outcomes (e.g., noting which Python version was used when a tool failed on the first attempt) to influence future responses. When a new chat begins, recent memories can be surfaced automatically or requested by the client to influence responses or tool usage.

Prerequisites:

• Node.js (and npm) installed on the host machine.
• Internet access to fetch packages.

Installation steps:

1. Verify Node.js and npm are installed:

   • node -v
   • npm -v

2. Install or update the MCP server package (via npx usage shown, package will be downloaded automatically):

   • Ensure network access to npm registry.

3. Run the MCP server (example using npx with the package name):

   • npx -y mcp-server-memory-file

4. Optional: If you prefer to pin a specific version, specify it in the command:

   • npx -y mcp-server-memory-file@1.0.0

5. Connect an MCP client to the running server as documented by your client; the server will expose the memory tools described in the README.

Tips and notes:

• This server uses a simple text file as the memory store; consider filesystem permissions and file path defaults when running in containerized or CI environments.
• If memory_delete/query returns no results, ensure your query matches or try a broader substring. memory_update is equivalent to delete followed by add.
• For tool usage reliability, you can prepend cue language in prompts to remind Claude when to store or retrieve memories (e.g., after a noteworthy interaction, prompt the model to store a memory).
• If you want to customize the memory store location or format, check the package options or configuration documentation provided by the mcp-server-memory-file package.
• Common issues: network restrictions blocking npm registry, conflicting Node versions, or file permission errors when writing the memory file.