mandaire.com For your life .app For your product .dev Architecture & trust .org

What Mandaire reads, and what it keeps

When you connect a source to Mandaire, what actually happens? What is read. What is stored today. Where we are building to. The straight answer to the question that blocks most beta signups.

Read →

Three things to build when your AI knows the person

Three patterns for using a personal knowledge graph in your product. Pre-meeting context enrichment. Per-recipient disclosure filtering. Context at query time, discarded after session. Each pattern is one MCP call.

Read →

What your AI forgets between sessions

Claude Code and Cursor reset every session. The problem is not the facts they forget. It is the decisions you made and why you ruled out the alternatives. Here is the fix.

Read →

Your app does not need to store your users' history.

The architecture where you read user context instead of owning it. What WWDC proved about which layer compounds and which layer you should stop rebuilding from scratch.

Read →

The brain is now a commodity. Your corpus is not.

Apple built the interface for swapping reasoning engines. WWDC confirmed what the architecture implied: the brain is rented, the corpus compounds. Here is what that means for the durable advantage in personal AI.

Read →

Apple's AI is Gemini. Mandaire is your life.

Apple showed its hand at WWDC. Siri now performs tasks using Gemini's reasoning and what it can see on your device. That is a real product. Here is what it is not.

Read →

Your AI queries are not private. Here is what to do about it.

When someone else queries your personal AI, what does it say? The question is not hypothetical. The answer depends on whether your AI has a disclosure policy or just instructions.

Read →

Why your AI needs a rules engine upstream of the model

Instructions are requests. A rules engine that runs before the model is a structural guarantee. The difference matters for anything you would regret reaching the wrong person.

Read →

Your AI has memory. What it doesn't have is a knowledge graph.

LLM memory accumulates. A knowledge graph resolves. Three structural gaps separate them, and no larger context window closes any of them.

Read →

Your AI is fluent. It just doesn't know you.

Frontier models are improving fast. The context they can see is not. Why the gap is structural, what we built to close it, and why ownership is the architecture that matters.

Read →

We built mandaire.app with Mandaire. Here is what we learned.

Not a demo. Months of directed work on the product you are reading about. Three calls that changed the architecture. A failure admission. What actually compounds.

Read →