I build the whole thing, from architecture, code, interface, release, to the 2am support tickets. My side projects have crossed 120,000 downloads with no paid marketing.
Selected Work
110,000+ downloads · 5,300+ stars · sponsored by Recall.ai
Open-source desktop AI assistant for interviews and calls. Screen + audio capture, always-on-top overlay, and live model responses — running locally on your machine.
The name started as a joke between friends. Then it blew up, got sponsored, and got cloned by a YC-backed company that shipped it as Glass without Git history or attribution.
I documented what happened in a Hacker News post that reached 947 points and 506 comments.
The company changed Glass's license to GPL, apologized to me and the open-source community, and credited Cheating Daddy on its project page.
Notion-style editor I built solo · led to editor work at Context.ai
Open-source document workspace in vanilla JS + Web Components. I built the editor core, block system, plugin architecture, themes, and templates.
LaTeX, charts, canvas, code, citations, custom plugins. Offline PWA with local autosave, Markdown / JSON / ZIP / PDF / DOCX export. Standalone open-source frontend, with backend services for sync, collaboration, and AI writing. GitHub
10,000+ downloads · 1,000+ stars · iOS + self-hosted dev infra
Control your dev machine from your phone, on your own infra. Expo client + Node CLI + Bun proxy + Rust PTY for real terminal sessions. No SSH setup.
The goal: make a phone feel like it's sitting inside your computer. QR pairing, dual WebSocket channels, files, git, processes, monitoring, terminals — plus a mobile UI for OpenCode / Codex agents that doesn't cram a desktop terminal onto a small screen.
Dev server on port 8000? Open localhost:8000 on your phone. No router changes, no ngrok. Custom mobile DevTools, end-to-end encrypted tunnels.
Lunel has reached 10,000+ downloads and 1,000+ GitHub stars without paid marketing. GitHub · App Store · Screenshots
How I Build
No hard line between product and engineering. I figure out what should exist, design the system, write the code, sweat the interaction details, deploy it, and fix what breaks in production.
I use AI heavily for research, prototyping, and debugging. It lets me try more ideas and get to working software faster — then I throw away the bad ones faster too.