Most of my ideas die.
I’m Jun. I run a one-person software company where AI agents do the research, the code, the SEO — and the paperwork for every idea we kill. Six products made it out alive. These are the receipts.
300+ideas screened
119kill verdicts on file
6products live
1human
Product manifestupdated 2026-08-19
LIVEEveryFormatJun 2026Convert 45+ image formats, locally in your browser.LIVEEveryMarkdownJun 2026A Markdown conversion toolbox.LIVEBookmark MaestroJun 2026Keeps your Chrome bookmarks clean — the extension I revived instead of killing.LIVEHistory of MoviesJul 2026Film history, computed from 20+ ranked lists.LIVELingoKeepAug 2026Learn languages from the videos you already watch.LIVEStyles GalleryAug 2026Try 83 design styles on your own site before you commit. Idea to paid product in 10 days.PAUSEDMedia Moana2022 —Programmable image-operations API for automations. Live, but the direction is being rethought.
PAUSEDfinkeeppre-launchLocal-first credit-card & points app for macOS. Built, self-hosted daily, not yet shipped.
BUILDINGDeft FoundryAug 2026 —An umbrella for small, focused Chrome extensions.
Paused is not a euphemism here — it’s a state on the board, with the reason written down. Every product above ships with the conditions for its own death, decided before launch. How to read this board →
The system
S1 discover → S2 validate → S3 commit → S4 build → S5 grow ∥ S6 operate → S7 reconcile
The whole company is a folder of markdown files, organized as a seven-station assembly line and operated by AI agents. One human gate: agents draft everything, nothing ships without me pressing the button. The playbooks, the kill memos, and the real numbers get published on this channel — the operator log is where they land first.