Recipes
Three declarative skin recipe JSON files built around different visual moods and reading-density tradeoffs.
Codex theme pack
The $9.99 starter pack gives you three data-only Codex skin recipes plus focused skin, pet and trust-boundary guides. It is built for people who want to learn faster without downloading executable theme code.
3recipe JSON files
$9.99one-time download
0executables included
What you are buying
The pack is not a secret installer and does not include gallery artwork. It packages reusable composition recipes, bilingual guidance and checklists so you can understand why a skin remains readable before sharing it.
Use it when you want to see practical settings for contrast, background framing, panel opacity, pet handoff and rights review. If you only need a single finished design, the custom service is more direct.
Inside the ZIP
Three declarative skin recipe JSON files built around different visual moods and reading-density tradeoffs.
A bilingual quick path from artwork selection to local preview, export and recovery notes.
A compact checklist for turning an owned image into a Codex pet brief and validating the resulting atlas.
Clear reminders for personal images, brand assets, fan art, stock photos and public redistribution.
How to use it
Complete Stripe checkout and unlock the protected ZIP after the order is verified.
Open the recipes and guides. Confirm there is no script, shell command, arbitrary CSS or bundled third-party artwork.
Use the free builder to recreate or adapt the settings with artwork you control.
Check Home, Task and Diff surfaces, then follow the install and restore notes.
Best buyer fit
Good fitYou want examples, bilingual guidance and reusable checklists for multiple skin experiments.
Bad fitYou expect celebrity artwork, a one-click installer, an official OpenAI product or guaranteed platform compatibility.
Upgrade pathOrder a custom skin when you already know the source image and want the finished result.
Free pathUse the builder and pets page without paying when you only need local experimentation.