Second Brain provides a seamless memory layer for every AI tool, allowing users to store information once and access it across platforms like Claude, ChatGPT, and Cursor. This self-hosted solution leverages Cloudflare's free tier to ensure quick setup and compatibility, making it easy to manage thoughts and insights across your AI interactions.
Second Brain — MCP Server on Cloudflare Workers
Unified Memory Layer for AI Tools
Streamline your interactions with artificial intelligence significantly by utilizing a single memory layer that integrates seamlessly with various AI tools, including Claude, ChatGPT, Cursor, and any other MCP-compatible clients. This self-hosted solution is designed to function on Cloudflare's free tier, allowing for rapid deployment and effortless memory management.
Key Features
- Persistent Memory Storage: With the Second Brain system, store information just once and access it across multiple applications. This prevents the common issue of forgetting details each time an AI session is initiated.
- Semantic Search: Powered by vector embeddings, this system allows for efficient semantic search capabilities, making it possible to retrieve relevant information using the meaning of the content, rather than simply keywords.
- Comprehensive Capture Options: Capture notes from various sources, including browsers, iOS devices, or scripts, and have your AI accesses relevant context at the beginning of every session.
Core Functionalities
The core functionality of Second Brain revolves around the following tools, each enabling specific actions:
| Tool | Parameters | Description |
|---|---|---|
remember | content (string), tags? (string[]), source? (string) | Store a note, with duplication checks in place. |
append | id (string), addition (string) | Append new information without creating a duplicate entry. |
recall | query (string), topK? (1–20, default 5), tag? (string) | Semantic vector search yielding relevant results. |
list_recent | n? (1–50, default 10), tag? (string) | Get a chronological list of recent entries. |
forget | id (string) | Delete an entry and its associated chunks. |
Dashboard Interface
Users can interact with the system via a built-in web UI, which provides three distinct views:
- Recall View — Search through memories semantically by asking natural language questions.
- Recent View — Access recent entries organized by date and filter by tags.
- Remember View — Create new memories directly with the option to use tags for better organization.
How It Works
Using Cloudflare Workers and an architecture that relies on storage systems like Cloudflare D1 and Vectorize, Second Brain operates through various capture sources including a dashboard, browser bookmarklet, iOS shortcuts, and API calls. Every note entered is transformed into a vector, enabling a true semantic search experience.
Getting Started
Initiating Second Brain is straightforward. A one-click deployment option enables users to set up their memory layer with minimal effort:
curl -X POST https://<your-worker-url>/capture \
-H "Authorization: Bearer YOUR_TOKEN" \
-H "Content-Type: application/json" \
-d '{"content": "Initial memory test", "source": "test"}'
Connect with AI Clients
Integrate this memory solution with popular AI platforms like Claude Desktop and Claude Code, ensuring a consistent memory experience across all interactions.
System Requirements
- A Cloudflare account (free tier supported)
- Basic familiarity with command line tools
This project enables stronger collaboration between humans and AI, as it retains valuable context and insights from past interactions, ultimately enhancing productivity and workflow.
No comments yet.
Sign in to be the first to comment.