Hardware Inventory

AI-powered hardware inventory management system. Take photos of your hardware, let AI identify and catalog it, then track market value over time with automated price lookups.

Features

• AI-Powered Intake - Upload photos of hardware items. AI vision models (Ollama local or Gemini cloud) automatically identify the category, brand, model, condition, and technical attributes.
• Human-in-the-Loop Review - Review AI results before cataloging. Correct any mistakes and re-analyze with AI using your corrections as context.
• Automated Pricing - When items are cataloged, AI estimates current used market value and original MSRP. Calculates depreciation automatically.
• Scheduled Price Refresh - Celery Beat refreshes all inventory prices on a configurable schedule (default: weekly).
• Price History Tracking - Track how item values change over time with historical price data and portfolio value charts.
• Search & Filter - Full-text search across all fields, faceted sidebar filters (category, condition, brand), and sortable table columns.
• Pluggable AI Providers - Configure multiple AI providers (Ollama for local, Gemini for cloud with web-grounded search). Switch between them from the settings page.

Quick Start

Prerequisites

• Docker and Docker Compose
• An AI vision model:
  • Local: Ollama with a vision model (e.g., llama3.2-vision)
  • Cloud: Google Gemini API key (enables web-grounded pricing)

Run with Docker

git clone https://github.com/yourusername/HardwareInventory.git
cd HardwareInventory
cp .env.example .env  # edit as needed
docker compose up -d --build

Open http://localhost:8080 in your browser.

Configure an AI Provider

1. Go to Settings in the nav bar
2. Add a provider:
   • Ollama (local): Set base URL to your Ollama instance (e.g., http://host.docker.internal:11434), model to a vision model like llama3.2-vision
   • Gemini (cloud): Paste your API key, model name gemini-2.0-flash (or similar)
3. Check "Set as Active Provider"
4. Click Save

Usage

1. Intake - Upload photos of hardware items
2. Review - Check AI analysis, correct if needed, click "Re-analyze with AI" for better results, then "Approve & Catalog"
3. Inventory - Browse, search, filter, and sort your cataloged items. View/edit details and pricing. Click "Refresh Price" for manual price updates.

Architecture

Frontend (Alpine.js + Tailwind)
    |
FastAPI (REST API)
    |
    ├── SQLAlchemy (SQLite)
    ├── Celery Worker (async AI processing + price lookups)
    ├── Celery Beat (scheduled price refresh)
    └── Redis (task broker)

Services (Docker Compose)

Service	Purpose
web	FastAPI app serving API + frontend (port 8080)
celery_worker	Async task processing (image analysis, price lookups)
celery_beat	Scheduled price refresh
redis	Celery message broker

API Endpoints

Method	Endpoint	Purpose
GET	/health	Health check
POST	/upload	Upload images for AI analysis
GET	/items/review	Items awaiting review
GET	/items/inventory	Cataloged inventory
PUT	/items/{id}/catalog	Approve and catalog an item
POST	/items/{id}/reanalyze	Re-analyze with user corrections
POST	/items/{id}/refresh-price	Manual price refresh
POST	/items/inventory/refresh-prices	Refresh all prices
DELETE	/items/{id}	Delete an item
GET	/categories	List categories
POST	/providers/	Add AI provider
GET	/providers/	List AI providers
GET	/items/{id}/history	Price history for item
GET	/portfolio/history	Portfolio value over time

Configuration

All configuration is via environment variables. See .env.example for the full list.

Variable	Default	Description
DATABASE_URL	sqlite:////app/data/inventory.db	Database connection string
CELERY_BROKER_URL	redis://redis:6379/0	Redis broker URL
ALLOWED_ORIGINS	http://localhost:8080	CORS allowed origins (comma-separated)
AI_TIMEOUT	120	AI request timeout in seconds
OLLAMA_URL	http://host.docker.internal:11434	Default Ollama URL
PRICE_REFRESH_DAYS	7	Skip items priced within this many days
PRICE_REFRESH_HOUR	3	Scheduled refresh hour (UTC)
PRICE_REFRESH_DOW	sunday	Scheduled refresh day of week

Development

Local Setup (without Docker)

cd backend
python -m venv venv
source venv/bin/activate  # or venv\Scripts\activate on Windows
pip install -r requirements.txt

# Start Redis (required for Celery)
docker run -d -p 6379:6379 redis:7-alpine

# Run the app
uvicorn backend.main:app --reload --port 8000

# In another terminal, start the Celery worker
celery -A backend.tasks.celery_app worker --loglevel=info

Project Structure

HardwareInventory/
├── backend/
│   ├── main.py           # FastAPI app and endpoints
│   ├── models.py          # SQLAlchemy ORM models
│   ├── schemas.py         # Pydantic request schemas
│   ├── database.py        # Database configuration
│   ├── ai_factory.py      # AI provider abstraction (Ollama, Gemini)
│   ├── pricing.py         # AI-powered price lookup
│   ├── tasks.py           # Celery tasks (image processing, pricing, scheduling)
│   └── requirements.txt   # Python dependencies
├── frontend/
│   └── index.html         # Single-page app (Alpine.js + Tailwind CSS)
├── data/sqlite/           # SQLite database (gitignored)
├── static/uploads/        # Uploaded images (gitignored)
├── docker-compose.yml     # Multi-service orchestration
├── Dockerfile             # Container image
├── .env.example           # Environment variable template
└── LICENSE                # MIT License

Security Considerations

This application is designed for personal/local use. If you plan to expose it to the internet, be aware of the following:

• No authentication - All endpoints are open. Add authentication (e.g., OAuth, API keys) before exposing publicly.
• API keys in database - AI provider API keys are stored in plaintext in SQLite. For production, consider using a secrets manager or environment variables instead.
• File uploads - Uploads are validated by extension and size (20MB max). Filenames are replaced with UUIDs to prevent path traversal.
• CORS - Restricted to configured origins by default. Update ALLOWED_ORIGINS in your .env file.
• Security headers - X-Content-Type-Options, X-Frame-Options, and Referrer-Policy headers are set on all responses.
• No rate limiting - Add a reverse proxy (nginx, Caddy) with rate limiting if exposing to the internet.

License

MIT