What is a personal context engine — and what makes it a Life OS?

What is a personal context engine?

Most AI conversations start cold. You paste a prompt, maybe the product remembers a few facts from earlier chats, and then you re-explain your calendar, your goals, your project and why Thursday workouts never stick.

A personal context engine flips that default. Instead of treating every session like an interview with a stranger, it keeps structuring the systems you choose to connect, so the AI can pull in the right context when you ask — or when a decision comes up on its own.

What the engine produces is a Life Knowledge Graph: the people, plans, habits, health signals and work items in your life, and how they relate. It is the difference between a pile of notes and a map you can ask questions of. The point is persistence and structure across domains — not just a longer chat history.

What is a Life OS?

A Life OS is an operating layer for the parts of your life — work, money, health and life — rather than one chat thread with a clever prompt.

In Decimal Life, those parts show up as modules: Health OS, Developer OS, Startup OS and Habits OS are examples, not a closed list. Every module is optional, and everything you switch on feeds the same Life graph. Each is a place where context turns into a decision and a next action, not another inbox of notes.

How is a Life OS different from a second brain?

A second brain — notes, a wiki, a PKM tool — is invaluable for capturing what you think. A Life OS goes further in two ways: it includes living systems (your calendar, your mail, your health signals, the work you are shipping) and it tries to help you decide and act, not only store.

Notes can feed a context engine. They are not the whole of it.

Why does AI still feel cold, even with memory?

Models are stateless, and product memory only partly fixes that: it is usually selective, built from what you typed, and thin next to a real week. It shows up as:

• re-explaining your goals every session;
• calendar constraints the model does not know about;
• sleep or energy that never makes it into the plan;
• project state stuck in GitHub or Linear while the chat stays generic.

That is a gap in the category, not a fault of any one chatbot. Chat memory is useful; it is just not the same thing as a graph built from the systems you live in. We go deeper on this in why your AI assistant keeps forgetting you.

How does a context engine work?

At a high level, four steps — and the first one is always yours.

1. You connect the sources you choose

   Today that is Gmail (read-only), your Google or Outlook calendar, and Apple or Google health data. Nothing is connected by default.

2. The engine builds and updates your Life Knowledge Graph

   People, plans, habits and signals become entities and relationships, kept current as the sources change.

3. Deci retrieves the context that matters

   When you ask a question, or when a module surfaces a decision, Deci reads from the graph rather than from chat history.

4. You get clarity and a next action

   Grounded in your real constraints — schedule, recovery, goals, the work you are shipping — instead of generic advice.

For example, if your calendar and your recovery both leave room, a health surface might suggest moving the workout you keep skipping — and ask you first, rather than quietly "optimising" your week. (An illustration of product behaviour, not medical advice.)

How does a context engine compare with chatbots and agents?

By criteria, not scores:

Swipe the table sideways →

Open-source context engines and Life OS projects exist too. Decimal Life's angle is connected life systems, Deci and domain modules together — currently in private beta, not a claim that every automation problem is already solved.

What is Decimal Life — and what isn't it?

How private is a personal context engine?

Trust is part of what the category means, so the controls matter as much as the features:

• Nothing is connected by default — you authorise each source.
• You can disconnect any app at any time.
• Export and delete controls work by domain, and derived context-engine data is treated with its source data on deletion.
• Inbox Radar reads Gmail; it never sends. Nothing leaves without you.

Decimal Life uses cloud infrastructure and model inference where the product needs them, so don't assume "local-only". Control lives in what you connect, what you export and what you delete. The full detail is in the Privacy Policy, and questions go to privacy@decimallife.com.

Who is a Life OS for?

Anyone weighing up personal AI, a Life OS or "AI that actually knows my life" — and tired of re-briefing it every session. A few examples, not a closed list:

• founders and builders living between a calendar and shipping tools, who need what matters tonight rather than a longer list;
• knowledge workers whose notes, mail and goals never talk to each other;
• anyone building habits or a health routine who wants context that checks before it assumes.

If your tools don't talk to each other, you are still the one joining them up. A personal context engine is built to take that job off you.