Available for software engineering roles - backend systems, production reliability, cloud deployment, and AI-assisted workflows.

From my build archiveActive

Backpocket OS

A local-first, privacy-driven Digital Twin for Australian sole traders. No cloud. No subscriptions. No data leaving your machine.

01

Repo notes

Primary language: Python

Technologies: Python, Streamlit, ChromaDB, Pydantic, Pytest

Topics: Not specified

Last updated: 2026-04-11T01:14:42Z

Stars: 0

Forks: 0

Status: Active

Visual notes

What the build looked like.

I have not added screenshots to this entry yet. The original README is still below if you want the less-polished version of the story.

BackPocket OS

The AIThatActually WorksFor You.

A local-first, privacy-driven Digital Twin for Australian sole traders. No cloud. No subscriptions. No data leaving your machine.


๐Ÿš€ The Pitch

The Problem: Australian sole traders are drowning in admin. Invoicing, BAS, ATO compliance, email triage โ€” it consumes 10+ hours every week. Existing "solutions" require cloud uploads, monthly subscriptions, and trust that big tech won't monetise your financial data.

The Solution: BackPocket OS runs entirely on your machine. Your business data never leaves your device. Using Agentic RAG and local large language models, it handles your business admin โ€” generating invoices, classifying emails against Fair Work guidelines, preparing BAS-ready reports โ€” while you stay compliant and in control.

Why Now:

  • Privacy is no longer optional โ€” GDPR, APPs, and client trust demand it
  • Local AI is finally fast enough to be practical (Gemma 4 E4B, Phi-3.5)
  • Sole traders are the backbone of Australian economy โ€” they deserve better tools

The Market: 1.2M+ sole traders in Australia. $3.2B spent annually on business admin tools. We're building the alternative that puts owners first.


๐Ÿ› ๏ธ Installation (Non-Dev Edition)

Prerequisites

  1. Python 3.12+

    # macOS
    brew install [email protected]
    
    # Ubuntu/Debian
    sudo apt update && sudo apt install python3.12 python3.12-venv
    
  2. Ollama (Local AI Engine)

    # macOS
    brew install ollama
    
    # Linux
    curl -fsSL https://ollama.ai/install.sh | sh
    

    Then pull your model:

    ollama pull gemma4:4b   # Recommended for speed
    # or
    ollama pull phi3.5      # If using Microsoft ecosystem
    
  3. Whisper (Voice transcription)

    ollama pull whisper
    

Setup

# Clone or extract this project
cd backpocket-os

# Create virtual environment
python3.12 -m venv venv
source venv/bin/activate  # macOS/Linux
# or
venv\Scripts\Activate  # Windows

# Install dependencies
pip install -r requirements.txt

# Start the app
streamlit run ui/app.py

That's it. Open http://localhost:8501 and start using your Digital Twin.


๐Ÿ” Security First: Zero-Cloud Architecture

BackPocket OS is designed with defense in depth. Here's how we protect you:

Data Flow Principles

LayerProtection
StorageAll data stays in ~/.backpocket/ โ€” local SQLite + ChromaDB. Never sent to external servers.
AI ProcessingOllama runs locally. Prompt inputs and outputs never leave localhost.
Vector DBChromaDB stores embeddings locally. No cloud sync.
NetworkApp operates offline. Internet only needed for initial dependency install.

Threat Model

ThreatMitigation
Data exfiltrationAll I/O bound to local filesystem. No external API calls in core flow.
Model poisoningVerify Ollama model checksums. Use GNUPG signatures where possible.
Memory scrapingProcess isolation via OS user permissions. No shared memory exposed.
Supply chainDependencies pinned in requirements.txt. Audit with pip audit.

What This Means

  • No cloud account required
  • No data ever transmitted (verified via network inspection)
  • You own your data โ€” export or delete anytime
  • Offline-capable โ€” works without internet

This isn't just "privacy-friendly." It's privacy-native.


๐Ÿ“‚ Project Structure

backpocket-os/
โ”œโ”€โ”€ core/                    # Business logic & models
โ”‚   โ”œโ”€โ”€ __init__.py
โ”‚   โ”œโ”€โ”€ invoice_engine.py    # Invoice generation
โ”‚   โ”œโ”€โ”€ compliance.py        # ATO/Fair Work rules
โ”‚   โ””โ”€โ”€ email_triage.py     # Email classification
โ”œโ”€โ”€ agents/                  # AI Agent orchestration
โ”‚   โ”œโ”€โ”€ __init__.py
โ”‚   โ”œโ”€โ”€ orchestrator.py     # Agentic RAG controller
โ”‚   โ””โ”€โ”€ prompts.py         # System prompts
โ”œโ”€โ”€ tools/                  # Tool definitions
โ”‚   โ”œโ”€โ”€ __init__.py
โ”‚   โ”œโ”€โ”€ pdf_generator.py   # PDF output
โ”‚   โ””โ”€โ”€ pdf_extractor.py  # PDF ingestion
โ”œโ”€โ”€ data/                   # Data stores
โ”‚   โ”œโ”€โ”€ chromadb/         # Vector embeddings
โ”‚   โ”‚   โ””โ”€โ”€ index/
โ”‚   โ””โ”€โ”€ ingest/           # Source documents
โ”œโ”€โ”€ ui/                     # Streamlit frontend
โ”‚   โ””โ”€โ”€ app.py
โ”œโ”€โ”€ config/                 # Configuration
โ”‚   โ””โ”€โ”€ settings.py
โ”œโ”€โ”€ TASKS.md               # Non-tech contributor tasks
โ”œโ”€โ”€ ARCHITECTURE.md        # System diagram
โ””โ”€โ”€ requirements.txt

๐Ÿงช Running Tests

pytest tests/ -v

๐Ÿค Contributing

We need more than code. See TASKS.md for ways to contribute without writing a single line.


๐Ÿ“œ License

MIT License โ€” because open source benefits everyone.


Built in Australia, for Australians.

Next conversation

Let's make the next system less fragile.

Open to software engineering roles across full-stack systems, platform and reliability work, workflow automation, and applied AI. I value teams where I can keep learning while contributing to real systems and clear delivery outcomes.

Also open to freelance or contract work across full-stack builds, practical automation, technical SEO, and cloud delivery.