Claude Code

Claude Code has full MCP support and a hooks system that enables automatic memory recall and storage without relying on the LLM to call tools.

Integration Options

Method	What it does	Setup effort
MCP only	LLM-driven memory via tool calls	Minimal -- just add MCP config
MCP + Plugin (recommended)	Automatic recall/remember via hooks + LLM tools for explicit ops	Copy plugin files + MCP config

MCP Setup

Add to your MCP configuration (~/.claude/claude_code_config.json):

Local (no auth):

{
  "mcpServers": {
    "mnemory": {
      "type": "streamable-http",
      "url": "http://localhost:8050/mcp",
      "headers": {
        "X-Agent-Id": "claude-code"
      }
    }
  }
}

Production (with auth):

{
  "mcpServers": {
    "mnemory": {
      "type": "streamable-http",
      "url": "https://mem.example.com/mcp",
      "headers": {
        "Authorization": "Bearer your-api-key",
        "X-Agent-Id": "claude-code"
      }
    }
  }
}

With MCP_API_KEYS configured on the server, user_id and agent_id are resolved automatically. No system prompt changes needed -- Claude Code injects MCP server instructions into the LLM's context.

Plugin Setup (Recommended)

The plugin uses Claude Code's hooks system to automatically recall memories at session start and store new memories after each exchange. This is more reliable than relying on the LLM to call tools.

See integrations/claude-code/ for the plugin code and detailed setup instructions.

Quick Install

1. Set environment variables:

   export MNEMORY_URL=http://localhost:8050
   export MNEMORY_API_KEY=your-api-key        # if using MCP_API_KEYS
   

2. Install the plugin:

   # Copy to your Claude Code settings directory
   cp -r integrations/claude-code/.claude-plugin ~/.claude/
   cp integrations/claude-code/hooks/hooks.json ~/.claude/hooks.json
   cp -r integrations/claude-code/scripts ~/.claude/scripts
   chmod +x ~/.claude/scripts/*.sh
   

3. Add the MCP config above for explicit tool access (search, update, delete).

How the Plugin Works

Hook	Action
SessionStart	Calls /api/recall to fetch core memories + instructions
UserPromptSubmit	Injects recalled memories as additional context
Stop	Calls /api/remember with the last user + assistant exchange

The plugin handles automatic recall/remember. The MCP tools remain available for explicit operations (searching mid-conversation, deleting a memory, etc.).

How Memory Works

With INSTRUCTION_MODE=proactive (default), Claude Code will:

1. Load context at session start -- calls get_core_memories (MCP) or gets injected context (plugin)
2. Search before answering -- when you ask about past decisions, project context, or "why did we do X?"
3. Store important context -- architecture decisions, coding conventions, project structure

Tips

• Use project:<name> categories -- tag project-specific memories with project:myapp, project:backend, etc.
• Pin architecture decisions -- important decisions should be pinned so they load at every session start
• Artifacts for deep analysis -- save full bug investigation reports or architecture research as artifacts
• Context type for current work -- "Currently working on X" uses context type with 7-day auto-expiry
• Shared vs agent-scoped -- technical preferences and project facts should be shared (no agent_id) so they're available across all your tools

See the coding assistant system prompt guide for more tips on what gets remembered.