406 creatures, one pipeline, zero manual uploads
We needed to turn a spreadsheet of monster stats into nineteen-plus separate storefront listings — each one packaged, priced, screenshotted, and described — without doing it by hand every time. So we built the pipeline instead of the listings.
An n8n workflow reads a Google Sheet, filters creatures by type, assembles the matching art and data into a package, compresses it, uploads it to Drive, generates marketing screenshots automatically, and drafts the store description with AI — Gemini and Grok helping with copy, a human doing the editing. The kind of workflow that looks simple in a demo and finds new ways to break in production, including a data-handling bug in the compression step that silently corrupted output until we traced it byte by byte.
Result: 19+ commercial products live on itch.io and Fab, with consistent pricing and presentation across the whole catalog through one repeatable workflow.
Built a vision-assisted Unity debugging tool
Every Unity developer eventually hits the same wall: your AI coding assistant can read code, but it can't see your game running. Menus misaligned, animations stuttering, physics doing something weird — none of that shows up in a stack trace.
The workflow routes screenshots and recordings through an AI vision model, converts them into a structured report, and hands that context straight to a coding assistant that can now actually see what broke. It's in daily use in our own dev pipeline.
Built a store-ready studio launch pipeline
The studio site, press kit, devlog, product offer, update capture, and launch checklist were built as one connected pipeline so marketing work has somewhere useful to land before the Steam page is ready.
Result: A deployable business surface with a playtest list, fixed-scope service offer, press materials, and screenshot workflow ready for the first playable build.
This is the kind of problem we like: something that should already be automatic and isn't. If that's what you've got, get in touch.