← All work

Language / Product

Mixolinguist

A private on-device AI companion and garden for learning, reflection, and personal growth across Apple platforms.

Role
Product direction, system architecture and development
Project status
In development
Mixolinguist identity artwork: a botanical form joins letters and circuitry.
Artwork from the project archive.

A private AI garden that grows with you

Mixolinguist is a private, on-device AI companion designed for learning, reflection, creation, and personal growth. While language practice remains a deeply integrated capability, the application has evolved into a quiet, purposeful companion across everyday life domains.

Unlike cloud-tethered assistants, Mixolinguist is built ground-up around local execution, user sovereignty, and verifiable trust:

  • Local-first intelligence — Runs core language models, evaluation loops, and reflection on-device using Apple Foundation Models, native Apple Silicon MLX runtimes, and local inference engines, with zero remote telemetry or transcript logging.
  • Structured Quests & Practice Coach — Goal-oriented activities, prerequisite chains, and an adaptive practice coach that helps turn insights from conversation into sustained creative, linguistic, or mindful habits.
  • Calm botanical design — A restrained interface crafted with the Alnwick-inspired Garden Design System, featuring circadian solar appearance shifts that mirror astronomical daylight phases.
  • Multiplatform & spatial — Built natively in Swift 6 for macOS, iOS, iPadOS, and visionOS—including an interactive 3D Lexicon Grove in spatial reality.

My contribution

Product direction, system architecture, interface design, and Swift development across all Apple platform targets, with AI coding assistance.

Current state

The application is actively in development. This portfolio entry documents the ongoing architecture, design systems, and product thinking; it does not announce a commercial App Store release.

Tags: Swift 6 · On-Device AI · Apple Intelligence · MLX · macOS / iOS / visionOS


Interested in the thinking behind this project?

Get in touch ↗