01
One durable home
A private repository becomes the canonical layer for identity, working context, memory maps, and trackers. Every assistant reads from the same source.
Open Source · MIT Licensed
Give your AI context you own.
ContextOS is a public template for building a private, version-controlled home base that your AI assistants read first—durable context, living trackers, automations, and a daily capture loop, all under your control.
The Problem
Powerful assistants. Blank-slate conversations.
Most AI tools are excellent inside a conversation and forgetful across them. Your priorities, preferences, projects, people, decisions, and operating rules end up scattered across chats—or trapped inside one vendor. ContextOS turns that context into a durable asset: readable by people, portable across capable assistants, and owned by you.
The model is swappable.
Your context is the durable asset.The ContextOS premise
Your context is the durable asset.The ContextOS premise
The Architecture
Three ideas hold it together.
ContextOS is not another app or a subscription. It is an operating pattern, packaged so your AI can install and maintain it for you.
02
The AI owns the paperwork
Tasks, relationships, and daily notes become living surfaces the AI reads and writes. You talk; it files, updates, and keeps the system coherent.
03
No silent automations
Every scheduled loop is listed in one registry—or it does not run. The system stays inspectable, recoverable, and under human authority.
The One Instruction
You do not install ContextOS. Your AI does.
Open the repository in Claude Code, Codex, or another assistant that can read files and follow a protocol. Then give it one instruction:
Tell your AI“Read BOOTSTRAP.md in this repo and set this system up for me.”
Your assistant surveys your environment, interviews you, proposes a system map for approval, and creates a private context repository adapted to your tools and devices.
What You Get
A complete foundation, not a rigid stack.
The core works anywhere an AI can read files. Optional modules are offered only when they fit your platform and workflow.
✓
Identity and context layer
Behavior, preferences, operating rules, memory map, and environment-specific entry points.
✓
AI-maintained trackers
Templates for tasks, relationships, projects, and daily notes with machine-readable structure.
✓
Daily capture loop
An optional path from phone capture to processed daily notes and synchronized action items.
✓
Portable modules
Optional phone widgets, persistent memory, local drafting workers, skills, and tracker automation.
✓
Agent governance
Clear authority boundaries, handoff patterns, and safeguards against silent expansion of access.
✓
Open Knowledge Format
Core trackers use OKF-compatible metadata so other tools can understand them without a custom parser.
Privacy by Design
The template stays public. Your life stays private.
ContextOS separates the reusable architecture from the personal system it creates.
The public repository
Generic forever. It contains the installer protocol, architecture, templates, optional modules, and hard-won operational lessons—never your personal data.
Your private repository
Private forever. It contains your identity, priorities, trackers, and working context. Secrets stay outside version control in platform secret stores.
Build Yours
Stop reintroducing yourself to your AI.
Start with the public template. Let your assistant adapt the architecture to the way you actually work.
Explore ContextOS on GitHub ↗