Your panel can host more than game servers.
Four of the most wanted self-hosted apps, packaged as proper Pterodactyl eggs. No Docker knowledge needed: import the JSON, create a server, the app binds to its allocation automatically.
What's inside
Built like eggs should be
Requirements
A working Pterodactyl 1.x panel with Wings. AMD64 nodes. Ollama needs 4 GB+ RAM for small models and disk space for models. Vaultwarden needs an HTTPS reverse proxy for real (non-test) use - explained in the README.
Four of the most wanted self-hosted apps, packaged as proper Pterodactyl eggs. No Docker knowledge needed: import the JSON, create a server, the app binds to its allocation automatically.
What's inside
- Ollama - local AI models with an HTTP API (OpenAI-compatible endpoint included). Optional PULL_MODEL variable auto-downloads a model on first start. CPU inference out of the box.
- n8n - the workflow automation tool, visual editor on the allocation port, timezone and secure-cookie variables included.
- Vaultwarden - Bitwarden-compatible password manager, with signup toggle and admin-page token as variables.
- code-server - VS Code in the browser, password-protected, persistent workspace and extensions.
Built like eggs should be
- Actually tested: every install script was executed in its real install container, and every startup command was booted and verified against its online-detection line before release
- Standard yolks runtime images, official upstream downloads - no custom Docker images to trust
- Sensible variables with validation rules, clear install-log messages, per-app notes in the README (RAM sizing for AI models, HTTPS requirements for Vaultwarden, first-boot times)
- All app data lives inside the server folder - normal panel backups just work
Requirements
A working Pterodactyl 1.x panel with Wings. AMD64 nodes. Ollama needs 4 GB+ RAM for small models and disk space for models. Vaultwarden needs an HTTPS reverse proxy for real (non-test) use - explained in the README.
