Apple Platform Developer
Practical utility apps for iPhone, iPad, and Mac — built privacy-first and on-device, with no accounts and no servers holding your data.
Utilities for the things you actually manage.
Add timestamps to your Blink and other videos
Process your Blink security camera footage with accurate timestamps burned directly into the video. Perfect for evidence documentation and sharing with authorities.
Evidence management for homeowners
Organize security footage, photos, and documentation for insurance claims or legal matters. Keep all your evidence organized and easily accessible in one place.
Your personal weather station, understood
Live dashboards for the spaces you manage — greenhouse, server closet, home — from your personal weather station, evolving into an on-device AI analyst for the climate of the space you manage.
Quiet, private tools for the day-to-day of recovery.
Ask the Big Book anything
The AA Big Book on your iPhone — ask questions in plain language and find any passage instantly. Answers come from the text, not the internet. On-device AI means your questions never leave your phone.
Recovery-focused home and lock screen widgets
Keep your recovery journey front and center with daily readings right on your home and lock screens. Free widgets that help you stay focused on what matters most.
On-device semantic search, end to end. Universal apps — one purchase across iPhone, iPad & Mac.
Build semantic search corpora from your documents
Turn any PDF or EPUB into a deployable RAG corpus package — no terminal, no Python, no manual file management. Embeddings are generated entirely on-device via Core ML. Built for developers and content curators building knowledge-first apps.
Ask your documents anything
The companion reader for the CorpusKit ecosystem — universal for iPhone, iPad, and Mac. Query pre-built corpora from the free catalog, or load your own .corpus bundles from Files or iCloud. On-device search, no account, no cloud.
Desktop-class tools for the files you've accumulated.
Powerful file organization for macOS
Clean up years of accumulated files on your Mac. Find duplicates, create smart sorting rules, batch rename, and auto-organize — nothing moves until you approve it.
Writing about any of these? Email support@robroy.online — I can often send an App Store promo code.
Built and in testing — not yet on the App Store.
Coming Soon
Network management for the rest of us
Discover every machine on your network, monitor them with lightweight agents, and get push alerts when something goes wrong. SSH, SFTP, and iOS Shortcuts integration — built by a DevOps veteran for home labs and small fleets.
Coming Soon
Conversations with the classics
Talk with public-domain historical figures — Diogenes, Confucius, Mark Twain, Nietzsche and a Vedantic sage — with a visible tone control that changes how each one answers you.
Tools and experiments published on GitHub — the working notes behind the apps.
Wildlife camera-trap triage. Sorts and annotates trail camera images with Google's SpeciesNet classifier — no cloud API, no account. On a Mac it can add behavioural context with Apple Foundation Models, entirely on-device.
View on GitHub (opens in a new tab)Converts all-MiniLM-L6-v2 from PyTorch straight to a Core ML .mlpackage, skipping the ONNX middle step. Targets FLOAT16 on the Neural Engine and exports a Swift-ready vocabulary for WordPiece tokenization.
A rule-based file organizer for the command line — match files by extension or filename pattern, preview every move with a dry run, then apply. Ships a config generator and a terminal config editor. Shares its name and idea with the Mac app; this is the Python one.
View on GitHub (opens in a new tab)A portable SwiftUI design system in a single package — colours, gradients, typography, spacing, shadows, view modifiers and pre-built components — so a set of apps can share one look without copying styles between projects.
View on GitHub (opens in a new tab)Not released yet, and not public — but they're what the shipped apps are built on.
The on-device AI layer the Mac apps share. It puts Apple Foundation Models and a local Ollama instance behind one typed API, where callers ask for a capability tier rather than naming a model — so an app never hard-codes which engine it runs on. No cloud calls, no third-party dependencies, and every task degrades gracefully to a working manual path when no model is available.
A self-hostable private RAG service — document retrieval and language-model inference that run entirely inside an organisation's own boundary, whether that's a VPC, an on-premise network, or an air-gapped enclave, so sensitive material and the questions asked of it never leave. The default configuration makes zero outbound calls, and the service reports that state on its health endpoint rather than just claiming it.
Get in touchWith 40 years of software development experience spanning IBM, Disney, and beyond, I've specialized in DevOps and systems administration before transitioning into iOS development. I build practical utility apps that solve real-world problems — the kind of tools I wish existed when I needed them. Each app reflects my methodical approach to development and my focus on creating solutions that genuinely make life easier.
I also take on consulting work for iOS, Mac and on-device AI projects. See what I can help with.
