Reflect Memory — Cursor Plugin

Cross-agent memory for AI tools. Read, write, browse, and search your shared memory layer directly from Cursor.

Reflect Memory gives your AI tools a shared long-term memory. Anything saved from ChatGPT, Claude, Gemini, or Cursor is accessible everywhere. This plugin connects Cursor to that memory via MCP.

Setup

1. Get your API key

Sign up at reflectmemory.com and generate an API key from the dashboard.

2. Install the plugin

Install from the Cursor Marketplace or add the MCP server manually:

Via Cursor Settings:

1. Open Cursor Settings (Cmd/Ctrl + ,)
2. Go to Features → Model Context Protocol
3. Click Add new MCP server
4. Name: reflect-memory, Type: stdio
5. Command: npx, Args: ["-y", "reflect-memory-mcp"]
6. Set env REFLECT_MEMORY_API_KEY to your key

Via project config (.cursor/mcp.json):

{
  "mcpServers": {
    "reflect-memory": {
      "command": "npx",
      "args": ["-y", "reflect-memory-mcp"],
      "env": {
        "REFLECT_MEMORY_API_KEY": "your-key-here"
      }
    }
  }
}

3. Restart Cursor

Restart Cursor for the MCP server to load.

Tools

Tool	Description
get_latest_memory	Get the most recent memory. Supports optional tag filter.
get_memory_by_id	Retrieve a single memory by UUID.
browse_memories	List memory summaries with filtering and search.
get_memories_by_tag	Get full memory content by tag(s).
write_memory	Save a new memory with title, content, tags.

Usage

Ask Cursor naturally:

• "Pull the latest memory"
• "What's the latest from ChatGPT?"
• "Browse all my memories"
• "Search memories for authentication"
• "Save this to memory"
• "Write a memory about the new billing system"

How it works

Reflect Memory is a shared context layer. When you write a memory from Cursor, ChatGPT can read it — and vice versa. Memories are tagged, searchable, and versioned.

Requirements

• Node.js 18+
• A Reflect Memory API key

License

MIT