# AI use cases for Uebox (ueboxai/uebox)

Repo: [ueboxai/uebox](https://github.com/ueboxai/uebox) | Use cases: 4 | Page: https://dedupedai.com/use-cases/ueboxai-uebox/

UE Box is an open-source TypeScript desktop AI agent that operates the Unreal Editor directly, working on levels, Blueprints, materials, and C++.

This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL.

- Let Codex or Claude Code edit an Unreal project via MCP: https://dedupedai.com/use-cases/ueboxai-uebox/drive-unreal-editor-from-external-agent/
- Ask an AI chat questions about an Unreal level: https://dedupedai.com/use-cases/ueboxai-uebox/ask-questions-about-unreal-level/
- Fix Blueprint and C++ compile errors with a goal checklist: https://dedupedai.com/use-cases/ueboxai-uebox/fix-blueprint-and-cpp-compile-errors/
- Call UE Box tools from a script or CI job: https://dedupedai.com/use-cases/ueboxai-uebox/run-uebox-checks-from-script-or-ci/

## How to let Codex or Claude Code edit an Unreal project via MCP with Uebox

Repo: [ueboxai/uebox](https://github.com/ueboxai/uebox) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README

Connect an external coding agent to a running Unreal Editor so it can call engine tools.

**Who it is for:** An Unreal developer who already uses Codex, Claude Code, or Cursor and wants that agent to change levels, Blueprints, or C++ in the open project.

**The problem:** Switching between an AI coding assistant and the Unreal Editor means copying work by hand, and the assistant cannot see or edit engine assets.

**The outcome:** The external agent reads and writes engine assets through UE Box tools, and UE Box lists every new, changed, or deleted item for the round.

### Steps

1. Download the installer from the Releases page of the repo and install UE Box on Windows.
2. In UE Box, go to 项目库 (project library), choose 导入项目, select your .uproject, install the UnrealAgentLink plugin when asked, then click 打开项目.
3. Follow the MCP guide at https://uebox.ai/guide/mcp and copy the local token it requires.
4. Add UE Box as an MCP server in Codex, Claude Code, or Cursor using that token. The MCP connection is writable by default.
5. Ask the agent for one small change, then check the change list in UE Box and use the re-read option to compare the editor state.

### Prompt to paste into your agent

```text
Using the UE Box MCP server in the ueboxai/uebox repo, count the light actors in the level <level name> and report the count. Then list the tool calls used and confirm the editor project is still open with no unexpected changes.
```

### Caveat

The README says MCP only connects from the local machine and needs a token, so a remote or cloud agent cannot reach the editor this way.

Risk, credentials: The MCP connection is writable by default and needs a token, so keep that token private and only connect on the local machine, since the README says it only works from localhost.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/ueboxai-uebox/drive-unreal-editor-from-external-agent/

## How to ask an AI chat questions about an Unreal level with Uebox

Repo: [ueboxai/uebox](https://github.com/ueboxai/uebox) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README

Use the built-in chat to query a level, such as counting lights, without writing code.

**Who it is for:** A level designer or artist who needs a quick count or check of what is in a level, and has an AI model key ready.

**The problem:** Checking a level by hand for things like light count or actor layout is slow, especially in large projects.

**The outcome:** The chat answers questions about the open level using engine tools, and the user sees each tool call that produced the answer.

### Steps

1. Install UE Box from the Releases page on Windows.
2. Go to 设置 → 模型 → 添加服务商, enter your API key, then set 默认模型 → 对话 to that model. The model must support tool calling.
3. Go to 项目库 → 导入项目, select your .uproject, install the UnrealAgentLink plugin, then click 打开项目.
4. Open 侧边栏 → AI 对话 and ask a question such as "How many lights are in this level?"
5. Check the tool calls shown in the chat to confirm the answer came from the open level.

### Prompt to paste into your agent

```text
Using the ueboxai/uebox app with an open Unreal project, ask the AI chat how many light actors are in the current level and return the number. Verify the answer by checking that the tool call list shows a level query, not a guessed number.
```

### Caveat

The README says the chat needs a model that supports tool calling, so a chat-only model without tool support will not query the level.

Risk, credentials: Your model provider API key is stored encrypted on your machine, and requests go straight to the provider you set, so only add keys you are comfortable using with that provider.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/ueboxai-uebox/ask-questions-about-unreal-level/

## How to fix Blueprint and C++ compile errors with a goal checklist with Uebox

Repo: [ueboxai/uebox](https://github.com/ueboxai/uebox) | Difficulty: Advanced | Time: 1 hour | Basis: From the README

Run goal mode to repair compile errors and verify each requirement in the project.

**Who it is for:** A gameplay programmer or Blueprint developer who has a list of requirements and a project that will not compile cleanly.

**The problem:** Compile errors in Blueprints or C++ block testing, and checking each requirement by hand after a fix takes time.

**The outcome:** UE Box repairs compile errors, runs the game to check logs and screenshots, and keeps working until each checklist item passes.

### Steps

1. Open the project in UE Box as described in the quick start, using 项目库 → 导入项目 and 打开项目.
2. Write your requirements as a checklist, one line per item, in the chat.
3. Start the request with /goal and paste the checklist so UE Box checks each item in turn.
4. Review the 本轮改动 list after each round to see created, modified, and deleted items.
5. Use the PIE run output, including logs and screenshots, to confirm the fix in play.

### Prompt to paste into your agent

```text
In the ueboxai/uebox app with the Unreal project <project path> open, use /goal with this checklist: <requirements>. Report which items pass after the final round and list the files and assets changed. Verify by checking that the compile step reports no Blueprint or C++ errors.
```

### Caveat

The README says the playtest bot for automatic key testing is experimental, so do not rely on it alone to confirm gameplay.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/ueboxai-uebox/fix-blueprint-and-cpp-compile-errors/

## How to call UE Box tools from a script or CI job with Uebox

Repo: [ueboxai/uebox](https://github.com/ueboxai/uebox) | Difficulty: Advanced | Time: 1 hour | Basis: From the README

Use the Uebox command line to ask a question or call one engine tool from automation.

**Who it is for:** A build or tools engineer who wants a scripted, repeatable check on an Unreal project, such as a nightly report.

**The problem:** Engine checks run by hand do not repeat reliably, and scripts have no way to query an open project.

**The outcome:** A script gets an answer or a tool result from the open project, and the default read-only mode keeps the project unchanged.

### Steps

1. Complete the setup in the quick start so the project opens in UE Box with the UnrealAgentLink plugin installed.
2. Make sure the local connection token is available to the script, following https://uebox.ai/guide/cli.
3. Run `uebox ask` from your script with a question about the open project.
4. Run `uebox tools call` from your script to call one engine tool and capture its output.
5. Store the output as a build artifact and fail the job if the expected value is missing.

### Prompt to paste into your agent

```text
Use the uebox command line from the ueboxai/uebox repo to ask the open project <project name> how many light actors the level has, and output the answer as JSON. Verify the job by confirming the command ran read-only and returned a numeric value.
```

### Caveat

The README says the connection is only available on the local machine, so a CI runner must be the same machine that has the Unreal Editor open.

Risk, credentials: The command line connects with a local token, so keep the token out of shared logs and CI output.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/ueboxai-uebox/run-uebox-checks-from-script-or-ci/