Skip to content

swappy-ops/lifeos_webui

Repository files navigation

LifeOS

Your life already generates telemetry. LifeOS turns that telemetry into reflection.

A fully local AI cognition layer for behavioral tracking, reflective analysis, and context-aware daily intelligence powered by Ollama, FastAPI, and local memory persistence.

LifeOS is not a cloud productivity app. It is a private, self-hosted behavioral interface designed to observe patterns across routines, synthesize context over time, and generate adaptive insights entirely on your machine.

No subscriptions. No cloud storage. No surveillance.

Just local intelligence.


Core Philosophy

Most trackers stop at: counting visualizing reminding

LifeOS attempts something different:

contextual reasoning behavioral interpretation adaptive reflection temporal memory daily prioritization

Instead of simply storing metrics, the system builds continuity across time using rolling contextual memory windows generated from historical logs.

The result is closer to:

a personal cognitive interface

than a traditional productivity dashboard.


Features

Fully Local AI Reasoning

Powered entirely through Ollama using locally hosted LLMs.

Your: health data financial tracking workouts behavioral logs reflections

never leave your machine.


Stateful Behavioral Memory

LifeOS automatically compiles rolling context windows from previous daily logs.

This allows the system to: remember trends maintain continuity generate context-aware insights avoid isolated one-shot analysis


FastAPI Backend

Lightweight and resilient Python backend using: FastAPI Pydantic validation asynchronous routing structured response handling


Robust JSON Extraction

LLMs hallucinate formatting.

LifeOS compensates using: deterministic RegEx extraction fallback handling validation layers crash resilience

to ensure stable structured outputs.


Minimal Native-Like UI

A dependency-light dark mode interface designed to feel tactile and responsive.

Features: animated transitions modal workflows responsive mobile access dashboard-style telemetry input lightweight architecture


Persistent Markdown Logging

Daily outputs are automatically archived into:

logs/daily.md

This acts simultaneously as: human-readable journaling exportable behavioral history AI memory substrate


Tech Stack

Layer Technology
Frontend Vanilla HTML/CSS/JS
Backend FastAPI
Validation Pydantic
AI Runtime Ollama
LLM qwen3:8b
Storage Markdown Logs
Serving Uvicorn

Project Structure

lifeos_webui/
│
├── index.html
├── lifeos.html
├── main.py
├── logs/
│   └── daily.md
├── README.md
├── LifeOS-main.zip
└── mistral-7b-instruct-v0.2.Q4_K_M.gguf

Architecture Overview

┌──────────────────────┐
│     Frontend UI      │
│ HTML • CSS • JS      │
└──────────┬───────────┘
           │
           ▼
┌──────────────────────┐
│    FastAPI Backend   │
│ Validation & Routing │
└──────────┬───────────┘
           │
           ▼
┌──────────────────────┐
│  Context Builder     │
│ Rolling Memory Logs  │
└──────────┬───────────┘
           │
           ▼
┌──────────────────────┐
│      Ollama API      │
│      qwen3:8b        │
└──────────┬───────────┘
           │
           ▼
┌──────────────────────┐
│ Structured AI Output │
│ Reflection + Actions │
└──────────────────────┘

Setup & Installation

1. Clone the Repository

git clone https://github.com/swappy-ops/lifeos_webui.git
cd lifeos_webui

2. Install Ollama

Download and install Ollama:

Ollama

Verify installation:

ollama --version

3. Pull the Model

ollama pull qwen3:8b

Make sure the Ollama server is running in the background.


4. Install Python Dependencies

pip install fastapi uvicorn pydantic requests

Optional but recommended:

python -m venv venv
source venv/bin/activate

Windows:

venv\Scripts\activate

5. Start the Application

uvicorn main:app --host 0.0.0.0 --port 8000 --reload

Usage

Desktop

Open:

http://localhost:8000

Mobile Access

Because the server binds to:

0.0.0.0

you can access the dashboard from devices on the same network.

Example:

http://192.168.1.50:8000

Daily Workflow

  1. Enter: sleep calories steps spending workouts notes

  2. Add gym exercises dynamically

  3. Click:

Analyze Day
  1. LifeOS: builds memory context retrieves historical patterns queries the local LLM validates outputs generates structured analysis archives the result

Privacy

LifeOS is intentionally local-first.

No: cloud APIs analytics tracking telemetry uploads external databases

Your behavioral data remains fully under your control.


Current State

LifeOS is currently an experimental cognitive systems project exploring: local AI interfaces behavioral reflection systems temporal memory quantified self workflows AI-assisted self-regulation

The architecture is intentionally lightweight and rapidly iterable.


Planned Directions

Near-Term

SQLite persistence structured memory indexing visualization graphs semantic tagging streaming responses modular prompt templates


Mid-Term

embeddings-based retrieval longitudinal pattern analysis emotional trend classification anomaly detection adaptive recommendations session summarization


Long-Term

multimodal input voice journaling calendar integration financial categorization autonomous planning agents local vector memory predictive behavioral modeling


Screenshots

Add screenshots and gifs here.

Suggested additions: dashboard overview mobile interface AI analysis output telemetry workflow


Why This Exists

Modern productivity systems optimize for: engagement streaks dopamine loops

LifeOS explores a different direction:

software as reflection.

Not performance theater. Not hustle dashboards.

A quieter system that helps interpret continuity across behavior, thought, and time.


Contributing

Contributions, experiments, and architectural discussions are welcome.

Areas especially relevant: local AI systems memory architectures HCI and UX quantified self tooling behavioral analysis agentic workflows


License

MIT License


Author

Built by swappy-ops


Final Note

LifeOS began as a daily tracker.

It is slowly evolving into something closer to:

a personal cognition substrate.

And that distinction changes everything.

About

LifeOS is a private AI cognition layer that observes patterns across sleep, movement, spending, workouts, and reflection to generate adaptive daily intelligence locally via Ollama.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors