Your AI's Private Memory Layer
xNotePadAI is the only note-taking app with a public MCP server. Give Claude, Cursor, Kiro, or any AI agent persistent, encrypted memory they can actually use.
Today's AI agents are stateless. Claude doesn't remember your last conversation. ChatGPT's memory is shallow and controlled by OpenAI. Cursor forgets your architectural decisions between sessions.
xNotePadAI solves this. Your notebook becomes persistent, private AI memory — a place where agents store and retrieve knowledge on your behalf, encrypted and under your control.
How it works — one token, 13 tools, zero trust architecture
🤖
AI Agent
Claude · Cursor · Kiro
ChatGPT · Any MCP Client
🔑
xNotePadAI MCP
Bearer Token Auth
13 Tools · Rate Limited
🔐
Your Notes
Encrypted · Private
Confidential Excluded
Your agent authenticates with a Bearer token. The server validates it, scopes access to your tenant only, and never returns confidential sections. You can revoke access with one click.
Your MCP Token
Copy this into your AI agent's MCP configuration. Already registered and ready to use.
Enable AI in Settings first Show full mcp.json config
13 Tools Your Agent Can Use
Read, write, search, organise, and reason over your notes — all via JSON-RPC 2.0 over HTTPS. No SDK needed, no complex setup.
list_notes
Browse all notes
create_note
Save new knowledge
get_note
Read full content
update_note
Modify content
delete_note
Remove AI notes
search_notes
Semantic RAG search
ask_notes
AI answers from notes
get_versions
Version history
tag_note
Organise with tags
link_notes
Connect knowledge
merge_notes
Combine duplicates
archive_note
Safe removal
index_notes
Bulk vectorise
These aren't just CRUD operations. ask_notes uses Llama 3.3 70B + RAG to synthesise answers from your entire notebook. search_notes finds notes by meaning, not keywords. index_notes bulk-vectorises your notes for instant semantic search.
Why This Matters
Every AI interaction today is ephemeral. You tell Claude something important, close the chat, and it's gone. Your research, decisions, context — lost to the void.
With xNotePadAI as your AI's memory layer, every interaction can persist. Your coding agent remembers your architecture decisions. Your research assistant recalls papers you read last month. Your meeting bot saves summaries you can search later.
The difference between a smart AI and a useful AI is memory. xNotePadAI provides that memory — privately, encrypted, and entirely under your control.
What You Can Ask Your Agent
Once connected, your agent can do all of this in natural language:
"Save this research to my xNotePadAI"
"What did I decide about the project timeline?"
"Organise my notes — tag everything, link related ones"
"Summarise everything I've written about AI this week"
"Merge my two meeting notes into one brief"
The agent maps these natural language requests to the appropriate MCP tool calls automatically. You never have to think about the API — just tell your AI what you need.
Security by Design
Giving an AI agent access to your notes requires trust. Here's how we earn it:
🔒 Confidential sections excluded
Any text marked confidential is never returned by any MCP tool — no matter what the agent asks for
🗑️ Can't delete human notes
Agents can only delete notes they created. Your notes can be archived (recoverable) but never destroyed
🔑 Scoped tokens
Each token accesses only your notebook. Tokens are SHA-256 hashed — plaintext never stored on server
⚡ Rate limited
30 requests/minute per IP. Prevents abuse, runaway loops, and cost overruns from misbehaving agents
Revoke a token at any time from Settings and the agent loses access immediately. No waiting, no grace period. You're always in control.
Who Uses This
Developers — give your coding agent persistent context. Architecture decisions, API notes, debugging logs that survive between sessions.
Researchers — let your AI assistant save papers, summarise findings, and build a searchable knowledge base over months of work.
Knowledge workers — meeting decisions, project briefs, client notes. Ask your AI "what did we decide about X?" and get an answer sourced from your own records.
AI builders — need persistent user memory for your app? xNotePadAI MCP provides it with zero infrastructure. One endpoint, 13 tools, auth included.
Private AI Memory — What Makes This Different
Persistent Context
AI agents save research, decisions, and context to your notebook. Next session, they pick up where they left off without you re-explaining everything.
You Own the Memory
Unlike ChatGPT's memory or Claude's project knowledge, this data lives in YOUR encrypted notebook. You can read, edit, or delete anything the AI stored.
Cross-Agent Memory
Claude, Cursor, and other agents share the same memory store. Research saved by one agent is accessible to all — your knowledge stays unified.
Portable Knowledge
Switch AI providers without losing context. Your memory layer is independent of any single AI vendor — it works with any MCP-compatible agent.
Why not just use ChatGPT's memory feature?
ChatGPT's memory lives on OpenAI's servers, is opaque (you can't easily see or manage it), only works within ChatGPT, and OpenAI can read it. xNotePadAI's AI memory is encrypted, fully visible and editable by you, works across multiple AI agents, and is stored under your control — not theirs.