Thoughts on AI-powered note-taking, privacy-first design, knowledge management, and building tools that respect your data.
Run a WebMCP validator scan to see whether AI agents can discover and use your site. What the score means, which checks matter, and how we took xNotePadAI from 48 to 99.
A Chrome DevTools extension that lists and runs every WebMCP tool a page registers. How to install it and point it at xNotePadAI's in-browser tools.
If notes are AES-256 encrypted, how can AI search them? The honest answer, with architecture diagrams.
Six independent security layers: token auth, tenant isolation, sandboxing, annotations, rate limiting, confidential exclusion.
Implementation details: UUID-pair tokens, SHA-256 hashing, tenant scoping, multiple tokens, instant revocation.
xNotePadAI vs Notion vs Obsidian vs Evernote vs Apple Notes. Encryption, MCP, privacy, and AI features compared honestly.
Turn your personal knowledge base into AI-accessible memory. Capture, organise, retrieve — all encrypted.
MCP servers expose tools that AI agents can discover and use. Learn how they work and how to connect to one.
Both connect systems, but they serve different consumers. Learn when to use each and why AI agents prefer MCP.
MCP tools are individual functions an AI agent can discover and call. Learn about schemas, annotations, and discovery.
AI agents use MCP to interact with external data. See which agents support it and how the safety model works.
Step-by-step: one token, 13 tools, under a minute. Connect Claude, Cursor, or any MCP agent.
Connect Claude Desktop or Code to your private notes. Save research, search by meaning, ask questions — with full encryption.
RAG retrieves context for one question. AI memory persists knowledge across sessions. Learn how xNotePadAI combines both.
MCP lets AI agents interact with external tools and data. Learn how it works, why it matters, and how xNotePadAI uses it for AI memory.
AI note taking uses artificial intelligence to help you write, organise, search, and connect your notes. Learn how it works and why privacy matters.
AI memory gives agents persistent context between sessions. Learn how private AI memory works and how xNotePadAI provides it via MCP.
localStorage caps at 4.5 MB. Here's how we work within that constraint — compression, auto-archive, smart expiry — without compromising your privacy.
Claude forgets. ChatGPT's memory is owned by OpenAI. xNotePadAI gives your AI agent encrypted, persistent memory via MCP — 13 tools, zero tracking, under your control.
Most AI note apps modify your content behind the scenes. We built mandatory diff approval instead — every AI change is shown as a diff before anything happens. Here's why that matters for trust.
Encryption and AI seem contradictory — how can AI understand your notes if it can't read them? We explain the architecture: encrypted at rest, decrypted only in your browser, AI processes content client-side or via confidential-section exclusion.
MCP (Model Context Protocol) lets AI agents like Claude and Cursor read and write to your notebook. This turns xNotePadAI from a writing tool into a persistent memory layer for AI — privately, securely, under your control.
The "second brain" movement assumes your notes live on someone else's server. What if you could have knowledge graphs, semantic search, AI summaries, and automatic linking — all encrypted and owned by you?
Most sync systems force "Keep Local" or "Keep Remote" — one side always loses. We added "Save Both" which preserves both versions and tags the conflict. Here's the thinking behind it.
When your notebook detects that one note mentions another by title and suggests a [[link]], isolated notes become a connected knowledge graph. We explain how automatic relationship detection works and why it matters.
More articles coming soon. Follow our journey building a privacy-first AI knowledge platform.