Building terminal-first tools, agentic workflows, and self-hosted systems.
I build software that makes dense systems legible:
- terminal workflows that survive machine moves and long sessions
- self-hosted infrastructure that can be rebuilt and understood
- agentic tools that stay useful under real operator pressure
- software that is useful before it is fashionable
yggterm: Rust-first remote terminal workspace for long-lived operator and agent sessions, now growing into libyggterm: the terminal as an application platform, where an app launched in a terminal takes over the GUI surfacesychrome: the first libyggterm app; run it in a yggterm terminal and the viewport becomes a browser on that machine's networkcellulose: open-source spreadsheet and financial modeling workspace; compatibility-first, automation-nativeyggdrasil: Debian live-build based ZFS + LXC host operating system
The company side of the work (incorporating): Practice, a CFA/GMAT practice platform built for my own exam preparation, and Jyas, a research engine that computes classical Jyotish deterministically so its claims can be tested honestly. Both are closed products in development.
yggterm: terminal workspaceychrome: web viewport, libyggterm pilotpaper: local-first structured writing and planning workspacecellulose: spreadsheet and financial modeling workspaceyggdrasil: host operating systemyggcli: guided configuration front dooryggsync: sync orchestratoryggdocs: ecosystem docs
codex-litellm: maintained LiteLLM patchset for the Codex CLI. In maintenance mode; the economics moved my daily agent work elsewhere, but it stays alive for the Codex surface and can be revived when the utility returnscodex-session-tui: terminal-first session explorer and repair tool for Codex conversationsthunderbird-cli: terminal workflow around mail search and operations
runtime-environment-flags: tiny runtime flags for JavaScript environmentsdav-manager: DAV and contacts tooling
Outside the code itself, I am interested in:
- finance and management as systems problems
- public-interest documentation and case synthesis
- writing book-length material from lived observation rather than generic advice
- X:
@avikalpax
This page is intentionally simple. I would rather show a few coherent systems than present a crowded wall of claims.



