PinWright

How it works

What you need to know to trust PinWright with your project: how your assistant connects, where your data goes, what it can change, and how you stay in control.

Your assistant works in your editor

PinWright is a plugin that runs inside the Unreal Editor on your machine. Your AI assistant connects to it and can then read your project, create and edit assets, play the game in the editor, and report back. You keep using the assistant you already have. There is no new app and no account.

Your assistant only loads the parts of PinWright a task needs. That keeps its attention on your project, even though PinWright covers 1,300+ editor actions. See what it covers.

Where your data goes

The plugin makes no network calls of its own and sends nothing to PinWright. Your assistant is a separate program. It sends what it reads from your project to the model provider you chose, as it already does with your code. Which provider that is, and whether the model runs locally, is up to you.

What it can change

Keeping or undoing changes

Everything your assistant makes is a normal asset in your project. Open it in the editor, play it, and keep what works. Use your version control to review what changed and to roll back the rest.

For Blueprints, materials, animation graphs and a few other asset types, your assistant can also show you the asset as readable text. That makes a change easy to review and compare. When it is rebuilt from text, the logic stays the same: nodes, connections and values. Layout details such as comment boxes, node colors and reroute knots are not kept.

Some tasks take a while

Most requests finish in seconds. A few take longer: snapshotting a large project, building lighting, or sweeping a big content folder can run from a minute to tens of minutes. Your assistant starts these, waits for them, and tells you when they finish, so you don’t have to repeat the request. You can keep working in the editor meanwhile.

Several projects at once

Each project folder gets its own connection port automatically, so you can run several projects side by side. Two editors of the same project would share a port, so only the first one accepts your assistant. If your team wants one fixed port, for example to share one assistant config, see Install and Troubleshooting.