MCP integration guide

Connect your memories to AI coding assistants.

What is MCP?

The Model Context Protocol (MCP) is an open standard that lets AI assistants access external tools and data. With SecondBrain's MCP integration, your personal memories become accessible context for AI workflows, right in your IDE.

This means Claude Code, Cursor or VS Code can query your memories, store new ones, and use your personal knowledge base while coding.

Prerequisites

Premium is required for the custom backend and MCP integration. Local capture and free iCloud sync work without this setup. Connected MCP clients can access backend memories under their own provider policies. Read the data handling details.

A deployed shodh-cloudflare backend
A supported Node.js LTS release installed
An MCP-compatible client (Claude Desktop, Claude Code, Cursor, etc.)

Quick setup

Step 1: Clone and install

git clone https://github.com/doobidoo/shodh-cloudflare.git
cd shodh-cloudflare
npm install

Step 2: Auto-configure

./scripts/setup-client.sh

The script asks for your Worker URL and API key, then configures Claude Desktop.

Step 3: Manual configuration (alternative)

All clients use the same server entry. Add it to the configuration file of your client:

Claude Desktop: ~/Library/Application Support/Claude/claude_desktop_config.json
Claude Code: .claude/mcp.json in your project or ~/.claude/mcp.json globally
Cursor or VS Code: your MCP settings

{
  "mcpServers": {
    "shodh-memory": {
      "command": "node",
      "args": ["/path/to/shodh-cloudflare/mcp-bridge/index.js"],
      "env": {
        "SHODH_CLOUDFLARE_URL": "https://shodh-memory.YOUR-SUBDOMAIN.workers.dev",
        "SHODH_CLOUDFLARE_API_KEY": "YOUR_API_KEY"
      }
    }
  }
}

Step 4: Verify

./scripts/verify-installation.sh

Restart your MCP client. You should see the shodh-memory tools available.

Available MCP tools

The MCP bridge exposes 13 tools for managing your memories:

Tool	What it does
`remember`	Store a new memory with type, tags and emotional metadata
`recall`	Semantic search across all memories
`recall_by_tags`	Filter memories by specific tags
`proactive_context`	Surface relevant memories from conversation
`list_memories`	Browse all memories with pagination
`update_memory`	Edit memory content or metadata
`forget`	Delete a specific memory
`forget_by_tags`	Bulk delete by tag
`batch_remember`	Import up to 50 memories at once
`reinforce_memory`	Boost the quality score of a memory
`consolidate`	Apply decay to old, low-quality memories
`memory_stats`	Get storage statistics
`context_summary`	Condensed view of recent decisions and learnings

Use cases

Coding with context

"Recall what I learned about the API rate limiting issue last week." Your AI assistant pulls relevant memories while you code, giving it personal context it would not otherwise have.

Cross-device memory

Capture a thought on your iPhone while walking, then access it from Claude Code on your laptop. All memories sync through your Cloudflare backend.

Project documentation

"Remember that we decided to use WebSockets instead of polling for the real-time feature." Store architectural decisions as memories and retrieve them when the topic comes up again.

Learning journal

Capture learnings throughout the day via voice on your Watch. Later, ask your AI assistant to summarize what you learned this week.

The full ecosystem

SecondBrain Pro (iOS and Watch) + shodh-cloudflare (backend) + MCP bridge (IDE) = one personal knowledge system across your devices and development tools.

Memories captured on your phone are available in your coding environment, and insights from coding sessions can be queried from your watch. All powered by your own Cloudflare infrastructure.

Resources

Backend source on GitHub: full source, documentation, troubleshooting
Cloudflare setup guide: deploy your backend
Model Context Protocol: MCP specification
SecondBrain Pro screenshots: app screens and details