UE5 Editor Plugin — AI-Generated DataTables
DataTable Streamliner
An Unreal Engine 5 editor plugin that generates fully populated DataTable assets from natural language descriptions using a local Ollama LLM. Describe your data — "12 fantasy weapons with name, damage, weight, and rarity" — and get a complete DataTable with an AI-defined or existing struct schema.
Preview generated rows in a table before committing to an asset. Append to existing DataTables with overwrite or merge options. The plugin discovers all UScriptStruct types in your project, or you can let the AI define the schema from scratch.
Like Input Streamliner, runs entirely local via Ollama with no cloud dependency. Accessible from the Tools menu in the Unreal Editor.
The Problem
Game designers and developers spend significant time manually populating DataTables in Unreal Engine — item databases, enemy stat sheets, dialogue tables, loot pools. Each row requires creating a struct, defining fields, and hand-entering values. For prototyping, this is wasted effort: you need 50 placeholder weapons to test your inventory UI, not hand-crafted game design. And when the schema changes, you rebuild the table from scratch.
Design Challenge
DataTables in UE5 are typed — every row must conform to a UScriptStruct with specific field types (float, int, FString, enum, etc.). An LLM generating data needs to produce values that match these types exactly, or the asset won't compile. The plugin also needs to work with existing project structs the developer has already defined, not just AI-invented schemas.
Key Design Decisions
The plugin auto-discovers every UScriptStruct registered in the project, so developers can select their own data structures as the target schema. Alternatively, the LLM can define a new struct from the natural language description. A preview table shows all generated rows before committing — the developer sees exactly what they'll get and can regenerate if the data doesn't look right. Append mode with merge/overwrite conflict resolution lets you build large datasets iteratively rather than generating everything at once. Like Input Streamliner, it runs fully local via Ollama — no external dependencies.
Outcome
Open-source UE5 plugin (MIT license) that generates populated DataTable assets from plain English descriptions. Useful for rapid prototyping — describe '20 sci-fi weapons with name, damage, fire rate, ammo type, and rarity' and get a complete, correctly typed DataTable in seconds. Published on GitHub with native Slate editor UI.
System Architecture
Architecture
Natural Language Data Generation
Describe your data requirements conversationally — item tables, enemy stats, dialogue entries, loot pools — and the plugin generates fully populated rows with appropriate values.
AI-Defined or Existing Structs
Let the LLM create a schema from your description, or select any existing UScriptStruct in your project. The plugin discovers all registered struct types automatically.
Data Preview
Review all generated rows in a table before creating the asset. Verify column types, values, and row count before committing.
Append Mode
Add rows to existing DataTables with overwrite or merge conflict resolution. Useful for iteratively building large datasets.
Gallery
DataTable Streamliner — Prompt & Settings
DataTable Streamliner — Data Preview
DataTable Streamliner — Generated Assets
DataTable Streamliner — Struct Selection
DataTable Streamliner — Tools Menu
DataTable Streamliner — Append Mode