┌────────────────────AXIOM OS ● RUNNING─────────────────┐
Current State
Focus
0%
Intent
DEEP_WORK
Goal Align
0%
Active Apps
  • Chrome21m
  • VS Code48m
  • Terminal12m
POLICY ENGINE✓ Activity aligned with declared goal
SPECTER DAEMON● Running│OLLAMA● Running
└────────────────────────────────────────────────────────────┘

OBSERVE

Daemon hooks

MEASURE

State tracking

STORE

SQLite memory

INTERPRET

Ollama eval

INTERVENE

Policy action

Core Specifications

  • Cloud Deps

    0

  • Telemetry

    Specter (Go)

  • Memory

    SQLite

  • AI Layer

    Ollama

System Trace

kernel: initializing telemetry daemon (Specter)...

Specter: hooked into process monitor.

SQLite: memory layer online.

Ollama: warming up qwen2.5:3b...

WARN: UNKNOWN state detected.

POLICY: productivity < 0.40 AND intent DISTRACT.

> executing contextual roast...

01 // The Problem

Current productivity assistants rely on cloud LLMs (privacy concerns) and act generically rather than enforcing personal accountability based on actual system behavior.

02 // The Synthesis

A local-first, zero-cloud personal operating system that observes telemetry, understands structured goals, and intervenes contextually using local AI.


System Architecture

AXIOM OS is an architectural design for a personal OS layer. It features a Go telemetry daemon (Specter) to collect local signals, a SQLite memory layer, and an AI brain powered by Ollama (Qwen 2.5). The core principle separates deterministic system state (measurement, policy) from AI interpretation, ensuring interventions are based on factual data, not LLM hallucinations. It includes a confrontational personality layer for strict goal alignment.

Tri-Axis Evaluation Model

AXIOM evaluates whether an activity is productive in context, not merely whether it advances a declared career goal. Three independent axes define the behavioral state:

Axis I

Current Role Duties

Axis II

Personal Goal Alignment

Axis III

General / Wellbeing

Rules before models. Evidence > Labels.

Key Subsystems

1. Local-First

Zero-cloud architecture ensuring all behavioral telemetry stays on the local machine.

2. Telemetry

Go-based Specter daemon collecting file, shell, and browser activity.

3. Determinism

Strict separation of deterministic measurement/policy from AI reasoning and conversation.

4. Intervention

Policy-controlled accountability system capable of contextual roasts and Focus Scoring.

Frequently Asked Questions

What is AXIOM OS?

AXIOM OS is a local-first personal AI operating system designed to run on your own hardware without relying on the cloud.

Is it cloud-based?

No, AXIOM OS is completely zero-cloud. It runs locally to ensure maximum privacy and offline availability.

How does it collect telemetry?

It uses a custom Golang daemon called Specter to collect contextual telemetry locally.

Why does AXIOM use deterministic policies?

To ensure predictable behavior and prioritize your designated goals over probabilistic distractions.

Which LLMs does AXIOM OS support?

It primarily utilizes local LLMs running via Ollama.

Where is the data stored?

All memory and context data are stored locally in SQLite databases.

Can I use AXIOM OS offline?

Yes, because it is local-first, it is fully functional offline.

Status: concept // SYSTEM ONLINE // EOF