LifeOS: An AI Harness for Life and Work
/LifeOS
LifeOS is a layer you install into an existing AI harness — Claude Code by default — adding persistent memory, a bundled skill library, and your own goals and context instead of a blank session every time. Reach for it if you already live inside an AI harness daily and want it to remember who you are; skip it if you don't run a capable harness yet, since LifeOS has nothing to attach to.
What is the LifeOS AI Harness?
As a general-purpose AI harness add-on, LifeOS is designed to ascertain your identity, your interests, and your future direction. This understanding then forms the basis for assisting with diverse tasks, from developing applications to launching enterprises. Installation occurs as a single, self-contained skill, which incorporates a comprehensive library encompassing research, security, writing, and art. This is then layered upon an AI coding harness, which was constructed and verified using Claude Code.
The Challenge of Unstructured AI Use
An AI coding harness by itself starts every session blank: it has no memory of your goals, your ongoing projects, or the decisions you made last week, so you re-explain context every time and get generic output regardless of who you are. LifeOS's own FAQ frames this directly — the harness is 'the engine,' but everything that makes it work toward YOUR goals across sessions is the part LifeOS adds.
Core Capabilities and Skills
- ✓Persistent memory: the README's FAQ says your AI remembers past sessions, decisions, and learnings instead of starting from zero each time.
- ✓A bundled skill library — research, security, writing, art, and more — installed as one self-contained skill, browsable at ourlifeos.ai/skills.
- ✓Intelligent routing: saying something like 'research this' is meant to trigger the matching workflow automatically, per the README's FAQ.
- ✓Self-improvement: the README describes the system modifying itself based on what it learns.
- ✓Harness-agnostic core built on hooks, skills, context files, and agentic routing rather than one vendor's proprietary features, though Claude Code is the most-tested path today.
- ✓A recovery path built in — the installer detects an existing install and merges intelligently, and your USER/ customizations are never touched by an upgrade.
How to Install LifeOS
LifeOS installs itself: paste 'Read https://ourlifeos.ai/install and install LifeOS for me.' into a capable AI coding harness — Claude Code, Cursor, Codex, Hermes, or similar — and your AI reads the install page and walks the setup, asking permission before touching anything. On macOS/Linux with Claude Code specifically, there's a one-line terminal shortcut: curl -fsSL https://ourlifeos.ai/install.sh | bash. Either path needs a capable AI coding harness (LifeOS is built and run on Claude Code) and bun.
Why Use LifeOS Over a Raw AI Harness?
- ✓Persistent memory across sessions, per the README's FAQ, instead of a harness that forgets everything the moment you close it.
- ✓A pre-built skill library you don't have to author yourself — research, security, writing, art, and more, all installed as one bundle.
- ✓A stated recovery path (backups, git history, a merge-aware re-installer) rather than leaving upgrade safety entirely up to you.
- ✓MIT-licensed and free, with the core design meant to port to any capable harness rather than locking you into one vendor.
Current Scope and Future Roadmap
- △LifeOS has nothing to attach to on its own — you need a capable AI coding harness first, and Claude Code is the only path the README calls well-tested today.
- △Several features are still on the roadmap, not shipped: local model support (Ollama, llama.cpp), granular per-task model routing, remote/mobile access, outbound phone calling, and external notifications (Email, Discord, Telegram, Slack).
- △A private source tree is used to generate the public repo, per the README, so community pull requests are ported in with credit rather than merged directly — an unusual contribution flow to be aware of.
- △Install and setup happen through an AI agent reading and acting on install instructions, which is a different trust model than downloading and running a normal package.
LifeOS vs. Fabric and Other Approaches
Frequently Asked Questions
A raw AI harness gives you capability with no memory of you: LifeOS adds persistent memory of past sessions and decisions, a bundled skill library, your own context (goals, contacts, preferences), and routing that triggers the right workflow automatically, per the README's own FAQ.
LifeOS is harness-agnostic by design, built on hooks, skills, context files, and agentic routing rather than one vendor's features — but Claude Code is what the README says it's actually built and run on, making it the most-tested path today.
Fabric comprises a range of AI prompts designed for particular tasks, effectively guiding the input for AI. For your assistant's day-to-day operations, LifeOS provides the foundational structure, covering aspects such as memory, skills, routing, and context. The README describes them as complementary, and says many LifeOS users fold Fabric patterns into their own skills.
The README documents a recovery path: back up first with a one-line copy command, your USER/ customizations are never touched by the installer, settings merge instead of overwrite, everything is git-backed, and re-running the installer detects an existing setup and merges intelligently.
Best use cases
- •Keeping an AI coding harness working toward the same goals across many sessions instead of re-explaining context each time.
- •Routing a request like 'research this' straight into the matching bundled skill instead of writing a fresh prompt from scratch.
- •Layering a memory and skills system onto Claude Code specifically, since that's what LifeOS is built and most tested on today.
- •Recovering from a broken install or upgrade using the built-in backup, git history, and merge-aware re-installer instead of starting over.
How to use
Once installed, LifeOS is used through the same AI harness you already interact with — the README doesn't document a separate CLI or UI. Its own FAQ describes triggering the right bundled skill by phrasing a request naturally (saying something like 'research this'), with the DA drawing on the memory and context LifeOS maintains rather than you re-explaining it each session.
Who should try it — and who should skip
Reach for LifeOS if you already run Claude Code (or another capable AI coding harness) daily and want it to remember your goals and context between sessions instead of starting blank. Skip it if you don't yet run a capable harness, or if you just want a library of prompts to paste in — that's closer to what Fabric, the README's own comparison point, already does.
Related repositories
Curious whether LifeOS is right for you?
Let ChatGPT, Claude, or Perplexity look into it — click below and see what AI actually says about LifeOS.