ABOUT CODEXE

Most AI coding tools give you a snippet and leave the hard part - turning it into something that actually runs - to you. Codexe produces the whole project instead: source files, build descriptor, manifest and configuration, laid out the way the target platform expects.

For Java targets it goes one step further and compiles the project on the platform, so you find out whether the code builds before you download it rather than after.

Every one of these is a target you can select today, not a roadmap item:

• Bukkit, Spigot and Paper plugins, in Java or Kotlin
• Folia plugins with region-aware scheduling
• BungeeCord and Velocity proxy plugins
• Forge and NeoForge and Fabric mods
• Vanilla datapacks and Skript scripts
• Discord bots in discord.py or discord.js
• Static websites and React applications

Blockbench model generation also exists and is marked ALPHA in the product - it works, but it is not yet held to the same standard as the targets above.

Codexe is model-agnostic. The catalogue includes models from Anthropic, OpenAI, DeepSeek and xAI, and which ones you can select depends on your plan. Generation is billed in tokens rather than per project, so a small script costs a fraction of a large multi-file mod. The credits and models documentation explains the detail.

Code you generate is yours. Nothing in the output calls home, and nothing requires a Codexe licence at runtime.

Codexe is built and operated by Penguin Studios, a sole proprietorship registered in Italy. The same entity is the data controller named in the privacy policy.

Service data is hosted in the EU. The full detail of what is collected and why is in the privacy policy, and the terms that govern use of the platform are in the Terms of Service.

Support, security reports and general questions all have their own channel - see the contact page.