Second Brain
Your own memory for AI tools on Cloudflare: save once, recall by meaning in Claude, ChatGPT, Cursor and any MCP client
Medium risk
We rate an entry medium when the tool runs code, makes network calls or reads project files. Check what exactly it does before installing.
Why this level
- Stores personal notes and decisions in a deployed service
- Requires keys and access to a Cloudflare account
Install
Manual install
npm install
npm run vectors:create
npm run deployDeploys a Cloudflare Worker to your account. Then connect MCP via the Connect to AI Clients guide.
This is third-party code. Review the repository files before installing.
What it does
Second Brain is one memory that deploys as a Cloudflare Worker in your own account and is available to any MCP-compatible AI tool. It saves decisions, preferences and notes, finds them by meaning rather than exact words, detects duplicates and relationships and ranks by time. Capture comes from MCP clients, a CLI, a browser extension, Obsidian, Notion, calendars and email. Data, vectors and keys stay in your Cloudflare account, and a team edition adds personal and shared layers.
Who it is for. For people who work across several AI tools and do not want to re-explain context in each.
Good fit when
- You repeat the same projects and preferences across AI apps
- You want memory in your own account, not someone else's cloud
- You need a shared memory layer for a team
Not a fit when
- You are not ready to deploy and maintain a Cloudflare Worker
- Local per-project memory is enough
Example request
Remember that in project Atlas we dropped Redis for a queue, and remind me next time we discuss infrastructureLimitations
You need a Cloudflare account; the memory runs on Workers with D1, Vectorize, Workers AI and KV. If Vectorize is down, capture and keyword recall keep working and semantic search recovers later. The shipped embedding models read English best; other languages need the multilingual mode in settings.
How to disable. Remove the MCP entry from the client config, and delete the Worker and its resources in the Cloudflare dashboard.
MCP
- Transport
- http
- Authentication
- API key
| Environment variables | |
|---|---|
| AUTH_TOKEN required, secret | Access token for your Worker, used in the Authorization header. |
Security check
- Stores personal notes and decisions in a deployed service
- Requires keys and access to a Cloudflare account
README in short
The README describes Second Brain as one memory for AI tools, deployed as a Cloudflare Worker in the user's account. Recall works by meaning, with classification, duplicate detection, relationships and time-aware ranking. Capture comes from MCP clients, CLI, browser, Obsidian, Notion, calendars, email and iOS shortcuts. Version v3 adds a team edition with personal and shared layers, per-person auth and hardened tenant isolation. Memory tools include remember, recall, append, link and forget.
FAQ
Where is data stored?
In your Cloudflare account: memories, vectors and keys live on your own Workers resources, outside anyone else's cloud.
Can a team use it?
Yes. In the team edition each person has a private layer, and the shared layer only holds what is deliberately shared, keeping author and permissions.
Related
A self-hosted knowledge base with block-level references and a built-in MCP server for connecting AI agents to your notes
A CLI for every Google Workspace API with JSON output and agent skills: Drive, Gmail, Calendar, Sheets and more
Local search over Markdown notes, docs and meeting transcripts: keywords, semantic search and reranking, with an MCP server
A task manager for AI-driven development: breaks a PRD into dependent tasks and guides the agent through them via MCP or CLI