NeoStack AI

https://docs.neostack.dev
Admin
NeoStack AI is an Unreal Engine plugin that integrates AI coding agents directly into the editor to...
Tokens:9,715
Snippets:196
Trust Score:4.8
Update:1 month ago
Show doc for...
Tokens:
Raw
### Install Copilot CLI

Source: https://docs.neostack.dev/clis/copilot

Install the Copilot CLI using various package managers or the provided install script.

```bash
npm install -g @github/copilot
```

```bash
brew install copilot-cli
```

```powershell
winget install GitHub.Copilot
```

```bash
curl -fsSL https://gh.io/copilot-install | bash
```

--------------------------------

### Initialize and Configure Claude Code

Source: https://docs.neostack.dev/clis/claude-code

Commands to start the Claude Code interface and install the NeoStack plugin.

```bash
claude
```

```text
/plugin marketplace add betidestu/neostack-plugins
/plugin install neostack@neostack-plugins
```

--------------------------------

### Install Gemini CLI

Source: https://docs.neostack.dev/clis/gemini

Install the Gemini CLI globally using npm.

```bash
npm install -g @google/gemini-cli
```

--------------------------------

### Start llama-server

Source: https://docs.neostack.dev/local-llms/llama-cpp

Launch the server with a specified GGUF model and context window size.

```bash
llama-server -m models/model.gguf -c 8192
```

--------------------------------

### Start vLLM Server

Source: https://docs.neostack.dev/local-llms/vllm

Launch the vLLM server with a specified model to begin serving inference requests.

```bash
vllm serve Qwen/Qwen2.5-Coder-7B-Instruct
```

--------------------------------

### Manual Installation Directory Structure

Source: https://docs.neostack.dev/getting-started/installation

The required folder structure for manual plugin installation within an Unreal Engine project.

```text
YourProject/
└── Plugins/
   └── NeoStackAI/
       ├── NeoStackAI.uplugin
        └── Source/
```

--------------------------------

### Install Codex CLI

Source: https://docs.neostack.dev/clis/chatgpt-codex

Install the Codex CLI globally using npm.

```bash
npm install -g @openai/codex
```

--------------------------------

### Prompt for AI Behavior Tree

Source: https://docs.neostack.dev/getting-started/quick-start

Example prompt for generating a Behavior Tree with specific logic sequences.

```text
Create a Behavior Tree called BT_PatrolAndChase with:
- Selector at root
- Sequence for chasing (check if player visible, move to player)
- Sequence for patrolling (move to random point, wait 2 seconds)
```

--------------------------------

### Prompt for Blueprint Actor Creation

Source: https://docs.neostack.dev/getting-started/quick-start

Example prompt for generating a Blueprint Actor with specific components and variables.

```text
Create a Blueprint Actor called BP_Coin with:
- A StaticMeshComponent called Mesh
- A SphereComponent called CollisionSphere
- A float variable called PointValue set to 10
```

--------------------------------

### Prompt for Blueprint Modification

Source: https://docs.neostack.dev/getting-started/quick-start

Example prompt for modifying an existing Blueprint by adding components and setting properties.

```text
Open BP_PlayerCharacter and:
- Add a SpringArm component attached to root
- Add a Camera component attached to the spring arm
- Set SpringArmLength to 300
```

--------------------------------

### Prompt for Health System Component

Source: https://docs.neostack.dev/getting-started/quick-start

Example prompt for generating a Blueprint Component with health variables and event dispatchers.

```text
Create a Blueprint Component called BPC_HealthComponent with:
- Float variable MaxHealth = 100
- Float variable CurrentHealth = 100
- Event dispatcher OnHealthChanged
- Event dispatcher OnDeath
- Function TakeDamage(float Amount) that reduces CurrentHealth and broadcasts events
```

--------------------------------

### Install Claude Code CLI

Source: https://docs.neostack.dev/clis/claude-code

Commands to install the Claude Code CLI on various operating systems.

```bash
curl -fsSL https://claude.ai/install.sh | bash
```

```powershell
irm https://claude.ai/install.ps1 | iex
```

```bash
brew install --cask claude-code
```

```powershell
winget install Anthropic.ClaudeCode
```

--------------------------------

### Prompting NeoStack AI for Blueprint Creation

Source: https://docs.neostack.dev/

Example prompt for generating a health pickup Blueprint with specific components and variables.

```text
Create a Blueprint called BP_HealthPickup that:
- Has a StaticMesh component for visuals
- Has a SphereCollision for overlap detection
- Has a float variable called HealAmount defaulting to 25
```

--------------------------------

### Vercel Model ID Examples

Source: https://docs.neostack.dev/api-providers/vercel-ai-gateway

Use these formats for model identifiers when selecting a model in the chat interface.

```txt
openai/gpt-5.4
anthropic/claude-sonnet-4.6
xai/grok-4.1-fast-reasoning
```

--------------------------------

### View Tool Audit Log Format

Source: https://docs.neostack.dev/troubleshooting

Example of the log structure recording AI actions, including timestamps, status, operations, and specific asset changes.

```text
[2026-02-08 14:23:01] OK   | edit_blueprint   | BP_Player | + Variable: Health (Float)
[2026-02-08 14:23:02] OK   | configure_asset  | BP_Player | bReplicates: false -> true
[2026-02-08 14:23:03] FAIL | edit_blueprint   | BP_Enemy  | Function 'OnDeath' not found
```

--------------------------------

### Launch Claude Code with Specific Project Directory

Source: https://docs.neostack.dev/clis/claude-code

Launch the CLI while explicitly defining the Unreal project directory.

```bash
NEOSTACK_PROJECT_DIR="/abs/path/to/MyProject" claude
```

--------------------------------

### Set Base URL for Qwen API

Source: https://docs.neostack.dev/api-providers/qwen

Use these base URLs in the provider settings to connect to the DashScope service.

```txt
https://dashscope.aliyuncs.com/compatible-mode/v1
```

```txt
https://dashscope-intl.aliyuncs.com/compatible-mode/v1
```

--------------------------------

### Authenticate Copilot CLI

Source: https://docs.neostack.dev/clis/copilot

Run the login command to authenticate the CLI before using it within the plugin.

```bash
copilot login
```

--------------------------------

### Make Bridge Executable

Source: https://docs.neostack.dev/clis/chatgpt-codex

Set execution permissions for the downloaded bridge file on macOS or Linux.

```bash
chmod +x ~/neostack-bridge-macos-arm64
```

--------------------------------

### Manual Model Configuration

Source: https://docs.neostack.dev/api-providers/qwen

Add these model identifiers manually if they are not automatically discovered.

```txt
qwen-plus
qwen-max
qwen3.6-plus
```

--------------------------------

### Configure MCP Server

Source: https://docs.neostack.dev/clis/copilot

Add the NeoStack bridge to the Copilot MCP configuration file to enable standalone terminal usage.

```json
{
  "mcpServers": {
    "unreal": {
      "type": "local",
      "command": "/path/to/neostack-bridge-macos-arm64",
      "tools": ["*"]
    }
  }
}
```

--------------------------------

### Configure MCP server in VS Code

Source: https://docs.neostack.dev/code-editors/vscode-copilot

Create this configuration file in the .vscode folder of your Unreal project to enable the bridge.

```json
{
  "servers": {
    "unreal": {
      "type": "stdio",
      "command": "/path/to/neostack-bridge-macos-arm64"
    }
  }
}
```

--------------------------------

### Configure MCP server in Rider

Source: https://docs.neostack.dev/code-editors/jetbrains-rider

Add the NeoStack bridge binary path to the stdio MCP server configuration in Rider's AI settings.

```json
{
  "mcpServers": {
    "unreal": {
      "command": "/path/to/neostack-bridge-macos-arm64"
    }
  }
}
```

--------------------------------

### Set LM Studio Base URL

Source: https://docs.neostack.dev/local-llms/lm-studio

The default local endpoint for the LM Studio server.

```txt
http://localhost:1234/v1
```

--------------------------------

### Enable Debug Mode

Source: https://docs.neostack.dev/clis/gemini

Run the Gemini CLI with debug logging enabled.

```bash
gemini --log-level debug
```

--------------------------------

### Add MCP Server

Source: https://docs.neostack.dev/clis/gemini

Register the NeoStack bridge as an MCP server for the Gemini CLI.

```bash
gemini mcp add unreal /path/to/neostack-bridge-macos-arm64
```

--------------------------------

### Configure DeepSeek Models

Source: https://docs.neostack.dev/api-providers/deepseek

List of model identifiers to be added manually if auto-discovery fails.

```txt
deepseek-chat
deepseek-reasoner
```

--------------------------------

### Configure Visual Studio AI extension for NeoStack

Source: https://docs.neostack.dev/code-editors/visual-studio

Add the NeoStack bridge as a stdio MCP server in your Visual Studio AI extension configuration file.

```json
{
  "servers": {
    "unreal": {
      "type": "stdio",
      "command": "C:\\path\\to\\neostack-bridge-win64.exe"
    }
  }
}
```

--------------------------------

### Reset MCP Configuration

Source: https://docs.neostack.dev/clis/gemini

Remove and re-add the unreal MCP server to resolve connection issues.

```bash
gemini mcp remove unreal
gemini mcp add unreal /path/to/neostack-bridge-macos-arm64
```

--------------------------------

### Configure SupportedTargetPlatforms in .uproject

Source: https://docs.neostack.dev/getting-started/installation

The JSON configuration required in the .uproject file to resolve packaging issues for non-desktop platforms.

```json
{
  "Name": "NeoStackAI",
  "Enabled": true,
  "SupportedTargetPlatforms": ["Win64", "Mac", "Linux"]
}
```

--------------------------------

### Configure LocalAI Base URL

Source: https://docs.neostack.dev/local-llms/localai

Set the base URL in NeoStack AI to point to your local LocalAI instance.

```txt
http://localhost:8080/v1
```

--------------------------------

### Pull multiple Ollama models

Source: https://docs.neostack.dev/local-llms/ollama

Commands to download several common models for use in NeoStack AI.

```bash
ollama pull llama3.2
ollama pull deepseek-r1
ollama pull qwen2.5-coder
```

--------------------------------

### Set Ollama context length

Source: https://docs.neostack.dev/local-llms/ollama

Environment variable configuration to increase the context window size for better tool handling.

```bash
# Windows (PowerShell): setx OLLAMA_CONTEXT_LENGTH 16384, then restart Ollama
export OLLAMA_CONTEXT_LENGTH=16384
```

--------------------------------

### Pull an Ollama model

Source: https://docs.neostack.dev/local-llms/ollama

Command to download a specific model to your local machine.

```bash
ollama pull qwen2.5-coder
```

--------------------------------

### Set Groq API Base URL

Source: https://docs.neostack.dev/api-providers/groq

Use this URL when configuring the custom provider in the NeoStack AI settings.

```txt
https://api.groq.com/openai/v1
```

--------------------------------

### Verify MCP Connection

Source: https://docs.neostack.dev/clis/claude-code

Command to verify that the NeoStack MCP server is connected.

```text
/mcp
```

--------------------------------

### Configure Ollama base URL

Source: https://docs.neostack.dev/local-llms/ollama

The default base URL for the OpenAI-compatible endpoint.

```txt
http://localhost:11434/v1
```

--------------------------------

### OpenAI-compatible chat endpoint path

Source: https://docs.neostack.dev/local-llms

The required endpoint path for NeoStack AI to communicate with local model servers.

```txt
/v1/chat/completions
```

--------------------------------

### Register Bridge with Codex

Source: https://docs.neostack.dev/clis/chatgpt-codex

Register the NeoStack bridge as an MCP tool within the Codex CLI.

```bash
codex mcp add unreal -- /path/to/the/bridge/you/downloaded
```

--------------------------------

### Configure Vercel AI Gateway Base URL

Source: https://docs.neostack.dev/api-providers/vercel-ai-gateway

Set this URL in the custom provider settings within NeoStack AI to route requests through the gateway.

```txt
https://ai-gateway.vercel.sh/v1
```

--------------------------------

### Set xAI Grok Base URL

Source: https://docs.neostack.dev/api-providers/xai-grok

The base URL required for configuring the xAI Grok custom provider in NeoStack AI settings.

```txt
https://api.x.ai/v1
```

--------------------------------

### Packaging Error Message

Source: https://docs.neostack.dev/getting-started/installation

The error message encountered when packaging for non-desktop platforms with an incorrectly configured plugin entry.

```text
NeoStackAI.uplugin is referenced via YourProject.uproject with a mismatched 'SupportedTargetPlatforms' field. This will cause problems in packaged builds, because the .uplugin file will not be staged.
```

--------------------------------

### vLLM API Endpoint

Source: https://docs.neostack.dev/local-llms/vllm

The default URL for the OpenAI-compatible API provided by vLLM.

```txt
http://localhost:8000/v1
```

```txt
http://localhost:8000/v1
```

--------------------------------

### Default Server Address

Source: https://docs.neostack.dev/local-llms/llama-cpp

The default local address where the server listens for requests.

```txt
http://127.0.0.1:8080
```

--------------------------------

### Verify Ollama Service Status

Source: https://docs.neostack.dev/community-faq

Use this command in your terminal to ensure the Ollama server is active and reachable.

```bash
ollama serve
```

--------------------------------

### Re-authenticate CLI Agents

Source: https://docs.neostack.dev/community-faq

Commands to trigger the authentication flow for specific CLI-based agents when sessions expire.

```bash
codex
```

```bash
gemini
```

```bash
claude auth login
```

--------------------------------

### Set DeepSeek Base URL

Source: https://docs.neostack.dev/api-providers/deepseek

The endpoint URL required for configuring the custom provider in NeoStack AI.

```txt
https://api.deepseek.com/v1
```

--------------------------------

### Update NeoStack Plugin

Source: https://docs.neostack.dev/clis/claude-code

Command to manually update the NeoStack plugin within Claude Code.

```text
/plugin update neostack@neostack-plugins
```

--------------------------------

### Uninstall NeoStack Plugin

Source: https://docs.neostack.dev/clis/claude-code

Command to remove the NeoStack plugin from Claude Code.

```text
/plugin uninstall neostack@neostack-plugins
```

--------------------------------

### NeoStack AI Base URL

Source: https://docs.neostack.dev/local-llms/llama-cpp

The endpoint URL required for configuring the custom provider in NeoStack AI.

```txt
http://127.0.0.1:8080/v1
```

--------------------------------

### Reset NeoStack npx caches in PowerShell

Source: https://docs.neostack.dev/community-faq

Use this command to manually clear corrupted npx caches if automatic retries fail. Ensure Unreal Engine is closed before execution.

```powershell
Remove-Item -Recurse -Force "$env:USERPROFILE\.neostack\npm-cache" -ErrorAction SilentlyContinue
```
Tokens:
Raw