Portfolio / Native macOS apps showcase
Everything runs on‑device on Apple Silicon. Your text, your voice, your machine. No cloud. No accounts. No surveillance.
Arunkumar V
Swift Developer · On-device ML
Two apps · 100% on-device
The Work
Make your writing sound like you again.
Humanizer detects AI-generated text with a locally-running MLX model on Apple Silicon, then rewrites it in a natural human voice using Apple's on-device Foundation Models (Apple Intelligence). Processing happens entirely on your Mac, your words never touch a server.
Requires macOS 14 Sonoma or later · Apple Silicon
Your conversations, understood deeply.
Voice_ is a conversation journal for macOS Tahoe. It records mic or system audio, live-transcribes with WhisperKit, performs speaker diarization, translates with cultural nuance, summarizes with context, and reads back via neural TTS. Every step runs entirely on-device.
Requires macOS 26 Tahoe · Apple Silicon
Zero network requests. Both apps are sandboxed macOS applications. No telemetry, no analytics, no accounts. Your data stays on your Mac, permanently.
A quick, one-time step, macOS remembers your choice.
These apps are built by an independent developer who does not yet hold a paid Apple Developer Program membership. macOS Gatekeeper may show a dialog reading “App can’t be opened because it is from an unidentified developer.” This is expected. Getting past it takes two clicks:
Locate the app in /Applications → right-click it → choose Open
In the dialog that appears, click Open
macOS stores your approval and will open the app normally on every subsequent launch. You only need to do this once. Source code for both apps is publicly available on GitHub, you are welcome to audit it before running anything.
About the Developer
Swift & SwiftUI
Native-first, platform-native
On-device ML
MLX, WhisperKit, Foundation Models
Privacy by design
Zero telemetry, zero cloud
I am Arunkumar V, an independent developer focused on Swift and the Apple platform ecosystem, specifically the intersection of native macOS app architecture, on-device machine learning with MLX, and the emerging capabilities unlocked by Apple Silicon. I believe software should feel inevitable: fast, private, and deeply integrated with the hardware it runs on.
Humanizer and Voice_ are not experiments, they are finished, usable tools I built to solve problems I actually had. Both demonstrate the same conviction: that privacy-first & on-device AI are not marketing claims but engineering choices made at the architecture level. Everything these apps do, they do on your Mac, using Apple’s own frameworks and the compute advantage of Apple Silicon.