
HPKV Memory
HTTP-SSEMCP server providing long-term memory capabilities for LLMs in Cursor IDE
MCP server providing long-term memory capabilities for LLMs in Cursor IDE
This repository contains the Cursor rule document for integrating the HPKV Memory MCP Server with Cursor IDE. The Memory MCP Server provides long-term memory capabilities for LLMs in Cursor, allowing AI assistants to remember conversations across sessions.
The Memory MCP Server implements the Model Context Protocol (MCP) to give LLMs true long-term memory. This rule document enables Cursor IDE to:
In short, it makes AI coding assistants in Cursor much more reliable by giving them access to persistent memory.
Create a free HPKV account and create an API Key in Dashboard.
Edit your mcp.json
file:
{ "mcpServers": { "hpkv-memory-server": { "command": "npx", "args": ["mcp-remote", "https://memory-mcp.hpkv.io/sse"] } } }
After adding the Memory MCP Server, you'll be prompted to login to your HPKV account. After that you can select the API key you generated.
Add the memory cursor rule to your project and set the rule type to Always
.
The memory_tool_usage_guide.mdc file contains the rules that instruct Cursor's AI how to properly use the memory tools. It includes:
In certain situations, mcp-remote
might have trouble refreshing your token and it consitantly generates new Client IDs that lead to the loop of trying to register a new Client ID, openning API Key selection page and goingback to generating a new Client ID. To fix this, disable the MCP server in Cursor, kill all mcp-remote
processes and clear the .mcp-auth
folder with a command similar to pkill -f mcp-remote && rm -rf ~/.mcp-auth
and then re-enable it in Cursor. This should fix any authentication issue you were experiencing.
For more information, check out our blog post about Memory MCP and the HPKV documentation.