Projects

Products, experiments, and systems built with curiosity and a bias toward making the next thing clearer.

Cartoon of Parth holding a wrench.

Product

  • RentMochi

    An intelligent operating layer for independent landlords.

    activeProduct
    • TypeScript
    • Next.js
    • Vercel
    • PostgreSQL
    • AI/LLM systems

    RentMochi is an attempt to rethink rental-property software around the landlord's actual information rather than around a collection of disconnected forms.

    It brings properties, tenants, documents, expenses, maintenance, communication, and decisions into one system, with AI helping users understand and act on that information.

    Project

Engineering

  • parthadhia.com

    My permanent corner of the internet.

    activeEngineering
    • Next.js
    • TypeScript
    • Vercel
    • GitHub
    • Cloudflare

    A personal identity, project, and infrastructure namespace designed to remain stable even as the services behind it change.

    The website is intentionally simple, mostly static, content-driven, and kept separate from private infrastructure sharing the same domain.

    ProjectRepository

Homelab

  • Homelab

    A private-by-default home infrastructure stack built around reliability, local services, DNS, automation, and recoverability.

    activeHomelab
    • Synology
    • Docker
    • Home Assistant
    • AdGuard Home
    • Unbound
    • Tailscale

    My homelab is less about collecting services and more about owning the infrastructure behind the things I depend on.

    The goal is a stack that remains understandable, recoverable, private by default, and boring when it needs to be boring.

Open source

  • Indic TTS Lexicon

    A pronunciation layer for Indian words inside English text-to-speech workflows.

    activeOpen source
    • TypeScript
    • JSON
    • IPA
    • SSML
    • TTS

    English TTS engines often badly mispronounce Indian names, Gujarati and Sanskrit terms, and devotional vocabulary.

    Indic TTS Lexicon provides structured pronunciation substitutions, IPA, and SSML so applications can produce more natural and respectful speech without modifying the underlying speech engine.

    Examples include terms such as Swaminarayan, satsang, mandir, and Indian names that conventional English phonetics handle poorly.

    ProjectRepository