Skip to main content


Inventories contain a dictionary of items and metadata, such as total weight or maximum capacity. These are an abstract data structure. Each inventory is uniquely identified by its name, which should be follow some conventions:

  1. The inventory name should be all-lowercase
  2. The inventory name should use dashes (-) to represent spaces

Data Structure

All inventories loaded in the session are cached by the server. All movement logic happens server-side, meaning there is no reliance on the client to make changes to the inventory structure. A typical inventory object follows this structure:

["rows"] = 5,
["max_weight"] = 100,
["weight"] = 32,
["name"] = "house-4893",
["items"] = {
["0"] = {
["name"] = "pistol",
["label"] = "Pistol",
["itemid"] = 453432689,
["weight"] = 12,
["stackable"] = false,
["quantity"] = 1,
["created"] = 1636664281,
["serial"] = "K1038384"
["2"] = {
["name"] = "pistolammo",
["label"] = "Pistol Ammo",
["itemid"] = 64,
["weight"] = 5,
["stackable"] = true,
["quantity"] = 4,
["created"] = 1636664362

Creating an Inventory

The inventory you might want to create depends on your use case. For instance, if the inventory is only needed for a current session, a temporary inventory should be created - that way unused inventories don't take up space in the database. Fortunately, the framework differentiates inventory types so this is quite simple to achieve.


A temporary inventory can be created using the inventory:createSecondaryInventoryTemp event.


A persistent inventory can be created using the inventory:createInventory event.