> ## Documentation Index
> Fetch the complete documentation index at: https://docs.baimoqilin.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Complete configuration

> Full MineClawd configuration reference with defaults, command keys, and behavior notes.

## Purpose

This page gives you the full MineClawd configuration model, including file settings, command keys, and restart behavior.

## Where settings live

* Global config file: `config/mineclawd.json5`
* Per-player chat visibility: `mineclawd/player-settings.json`
* Soul files and active soul state: `mineclawd/souls/` and `mineclawd/souls/.active/`

## Complete `mineclawd.json5` template

```json5 theme={null}
{
  provider: "OPENAI",
  endpoint: "https://api.openai.com/v1",
  apiKey: "",
  model: "gpt-5.2",
  summarizeModel: "gpt-5.2",

  vertexEndpoint: "https://aiplatform.googleapis.com/v1",
  vertexApiKey: "",
  vertexModel: "gemini-3.1-pro-preview",
  vertexSummarizeModel: "gemini-3-flash-preview",

  tavilyApiKey: "",

  debugMode: false,
  limitToolCalls: false,
  toolCallLimit: 16,
  systemPrompt: "",
  dynamicRegistryMode: "AUTO",

  // Client-only (not in server config file)
  enableGui: true
}
```

<Info>
  In file form, enum values are uppercase (`OPENAI`, `VERTEX_AI`, `AUTO`, `ENABLED`, `DISABLED`).

  In command form, values are lowercase (`openai`, `vertex-ai`, `auto`, `enabled`, `disabled`).
</Info>

## Key reference

| File key               | Command key              | Type    | Default                                | Notes                                                                                                  |
| ---------------------- | ------------------------ | ------- | -------------------------------------- | ------------------------------------------------------------------------------------------------------ |
| `provider`             | `provider`               | enum    | `OPENAI`                               | Provider selection: OpenAI-compatible or Vertex AI.                                                    |
| `endpoint`             | `endpoint`               | string  | `https://api.openai.com/v1`            | OpenAI-compatible base URL, no trailing slash required.                                                |
| `apiKey`               | `api-key`                | string  | empty                                  | OpenAI API key.                                                                                        |
| `model`                | `model`                  | string  | `gpt-5.2`                              | OpenAI chat model.                                                                                     |
| `summarizeModel`       | `summarize-model`        | string  | `gpt-5.2`                              | OpenAI model for first-turn session title generation.                                                  |
| `vertexEndpoint`       | `vertex-endpoint`        | string  | `https://aiplatform.googleapis.com/v1` | Vertex API base URL.                                                                                   |
| `vertexApiKey`         | `vertex-api-key`         | string  | empty                                  | Vertex API key.                                                                                        |
| `vertexModel`          | `vertex-model`           | string  | `gemini-3.1-pro-preview`               | Vertex model path or model name.                                                                       |
| `vertexSummarizeModel` | `vertex-summarize-model` | string  | `gemini-3-flash-preview`               | Vertex summarize model for session titles.                                                             |
| `tavilyApiKey`         | `tavily-api-key`         | string  | empty                                  | Tavily API key for web search. When set, enables the `search` tool.                                    |
| `debugMode`            | `debug-mode`             | boolean | `false`                                | Logs detailed request and tool activity.                                                               |
| `limitToolCalls`       | `limit-tool-calls`       | boolean | `false`                                | Enables per-request tool-loop cap.                                                                     |
| `toolCallLimit`        | `tool-call-limit`        | integer | `16`                                   | Allowed range is `1` to `20`.                                                                          |
| `systemPrompt`         | `system-prompt`          | string  | empty                                  | Empty means built-in system prompt. Use value `default` in command mode to reset.                      |
| `dynamicRegistryMode`  | `dynamic-registry-mode`  | enum    | `AUTO`                                 | Controls dynamic placeholder runtime mode. Restart required after change.                              |
| `enableGui`            | --                       | boolean | `true`                                 | Client-only. Show or hide the overlay GUI. Set through **Mod Menu** -> **MineClawd** -> **Configure**. |

## Per-player settings

### broadcast-requests-to

Command-only, per-player.

* Command key: `broadcast-requests-to`
* Allowed values: `self`, `all`, `ops`
* Stored in: `mineclawd/player-settings.json`
* Purpose: controls who sees prompt echo and task start or finish lines.

### assistivetouch

Toggle-only, per-player.

* Command: `/mineclawd assistivetouch`
* Stored in: `mineclawd/player-settings.json`
* Default: enabled (visible)
* Purpose: show or hide the floating AssistiveTouch orb without disabling the full GUI.

## Dynamic registry mode behavior

* `AUTO`: enabled in client runtime (single-player), disabled on dedicated servers.
* `ENABLED`: force enabled in all runtimes.
* `DISABLED`: fully disabled.

<Warning>
  If you force `ENABLED` on a dedicated server, joining clients must have MineClawd installed.
</Warning>

## Recommended baselines

### Baseline for first-time OpenAI setup

```bash theme={null}
/mineclawd config provider openai
/mineclawd config endpoint https://api.openai.com/v1
/mineclawd config api-key <your-key>
/mineclawd config model gpt-5.2
/mineclawd config summarize-model gpt-5.2
```

### Baseline for first-time Vertex setup

```bash theme={null}
/mineclawd config provider vertex-ai
/mineclawd config vertex-endpoint https://aiplatform.googleapis.com/v1
/mineclawd config vertex-api-key <your-key>
/mineclawd config vertex-model gemini-3.1-pro-preview
/mineclawd config vertex-summarize-model gemini-3-flash-preview
```

## GUI path

* In game: **Mod Menu** -> **MineClawd** -> **Configure**
* Command fallback: `/mineclawd config ...`
* Overlay menu: click the hamburger icon -> **Config**

If client GUI open fails, configure through command keys.

<Info>
  The `enableGui` setting controls the overlay window. Setting it to OFF does not affect the YACL config screen -- you can still open config through **Mod Menu** or `/mineclawd config`.
</Info>