Skip to content

Allow anonymous analytics?

Our cookieless analytics show us which features get used and store nothing on your device. We never sell your data or use it for advertising.

Privacy choices

Choose what ModDex may collect. You can change this anytime from Privacy choices in the footer.

Read the full privacy policy

Strictly necessary Required for the site to work: login sessions, security tokens, and abuse prevention. Always on and cannot be disabled. Always on
Analytics Cookieless analytics through PostHog's EU servers, used only to see which features get used. Nothing is stored on your device, and the anonymous visitor ID resets every day. Needs your consent because usage data goes to a third party. Never used for advertising or profiling. Off by default in the EU and UK.

Sever Agent

Quick rating

Community listing page, reviews may not be monitored by the author.

How listings work Are you the author?

Sever Agent

By youyiMC

No reviews yet

This mod lets your Minecraft server connect to AI service APIs

Mod Loaders
Minecraft
Shockbyte server hosting: 35% off your first server with code MODDEX35 (opens in a new tab)

About

Description

Sever Agent — AI Assistant for Minecraft Servers

An AI chatbot assistant for your Minecraft server. Just install it on the server (compatible with single-player / LAN / dedicated servers) — no client mod required.

Supported versions: Minecraft 1.21.1 · NeoForge

✨ Features

  • 🗨️ Chat with AI — players ask questions with /ai <message> or /ai ask <message>, and the answer is delivered back as a private message (/tell effect).
  • 🌐 Web Search — built-in presets that automatically declare web search to the connected model. Supports ChatGPT/OpenAI, DeepSeek, Zhipu GLM, Qwen (Tongyi), ERNIE (Wenxin), Doubao (Volcengine Ark) and more.
  • ⚙️ Hot-reload config — change everything at runtime via /ai config set <key> <value> or /ai reload, no server restart needed.
  • 🚫 Admin controls — ban/unban players from using the AI with /ai ban <player> / /ai unban <player>.
  • 🧠 Deep thinking support — optionally show the model's reasoning process (showThinking), if the model supports it.
  • 🔀 Smart queue & rate limiting — a shared server-wide FIFO queue with sliding-window rate limits and global cooldown; never floods your API key.
  • 📦 Full, non-streamed replies — waits for the complete response before sending; long answers are split automatically.
  • 🌍 Multi-language — Simplified/Traditional Chinese (HK & TW), Classical Chinese, English, Spanish, French, Arabic, Russian and Japanese.

📖 Commands

Command Description Permission
/ai <message> Ask the AI (same as /ai ask) Player
/ai ask <message> Ask the AI, reply via private message Player
/ai status Show API config, rate-limit & queue status Player
/ai reload Reload config from disk Admin (OP 2)
/ai config set <key> <value> Change config at runtime, applied immediately Admin (OP 2)
/ai ban <player> Ban a player from using the AI Admin (OP 2)
/ai unban <player> Unban a player Admin (OP 2)

⚙️ Configuration

Generated at config/severagent-server.toml on first run (NeoForge native TOML).

Key options (also changeable via /ai config set):

Key Description
apiKey Your API key (leave empty to disable AI)
baseUrl OpenAI-compatible endpoint URL
model Model name
webSearchEnabled Enable web search (default off)
webSearchMode Preset: chatgpt / deepseek / zhipu / qwen / ernie / doubao / search_field / search_enabled_field
windowMinutes Rate-limit window (only 1 / 5 / 10)
maxRequestsPerWindow Max requests per window
cooldownSeconds Global cooldown between requests
showThinking Show the model's thinking process
systemPrompt System prompt template ({player_name} placeholder)

🚀 Quick Start

/ai config set apiKey sk-xxxx
/ai config set webSearchEnabled true
/ai config set webSearchMode chatgpt
/ai 今天天气怎么样？

Note: Web search only works if the model/API you connect to actually supports it. DeepSeek's official API does not offer native web search — use it through a compatible gateway, or use the aggregator platform's preset (e.g. qwen) instead.

Sever Agent —— 我的世界服务器 AI 助手模组

为你的 Minecraft 服务器接入 AI 聊天助手。 仅需在服务端安装（兼容单人游戏 / 局域网开放 / 专用服务器），客户端无需安装。

支持版本：Minecraft 1.21.1 · NeoForge

✨ 功能特性

  • 🗨️ 与 AI 对话 —— 玩家用 /ai <内容> 或 /ai ask <内容> 提问，回答以私信（/tell 效果）回传。
  • 🌐 联网搜索 —— 内置主流大模型联网声明预设，自动生成对应平台的请求字段，支持 ChatGPT/OpenAI、DeepSeek、智谱 GLM、通义千问、文心一言、豆包（火山方舟） 等。
  • ⚙️ 配置热更新 —— 通过 /ai config set <键> <值> 或 /ai reload 运行时修改配置，无需重启服务器。
  • 🚫 管理员封禁 —— /ai ban <玩家> / /ai unban <玩家> 冻结/恢复玩家对 AI 的访问权限。
  • 🧠 深度思考支持 —— 可配置是否展示模型的思考过程（showThinking），适用于支持深度思考的模型。
  • 🔀 智能队列与限流 —— 全服共享 FIFO 请求队列，配合滑动时间窗口限流与全局冷却，避免 API 被并发打爆。
  • 📦 完整响应、非流式回传 —— 等待完整回复后再发送，绝不流式逐条截断；超长回复自动分片。
  • 🌍 多语言支持 —— 简体/繁体中文（香港、台湾）、文言文、英语、西班牙语、法语、阿拉伯语、俄语、日语。

📖 指令一览

指令 说明 权限
/ai <内容> 直接提问（等价于 /ai ask） 玩家
/ai ask <内容> 向 AI 提问，结果私信回传 玩家
/ai status 查看 API 配置、限流与队列状态 玩家
/ai reload 从磁盘热重载配置 管理员（OP 2）
/ai config set <键> <值> 运行时修改配置并立即生效 管理员（OP 2）
/ai ban <玩家> 封禁玩家使用 AI 管理员（OP 2）
/ai unban <玩家> 解除封禁 管理员（OP 2）

⚙️ 配置说明

首次运行自动生成 config/severagent-server.toml（NeoForge 原生 TOML 配置），也可通过 /ai config set 修改：

键 说明
apiKey API 密钥（留空禁用 AI）
baseUrl 兼容 OpenAI 协议的接口地址
model 模型名称
webSearchEnabled 是否启用联网搜索（默认关闭）
webSearchMode 预设：chatgpt / deepseek / zhipu / qwen / ernie / doubao / search_field / search_enabled_field
windowMinutes 限流时间窗口（仅 1 / 5 / 10）
maxRequestsPerWindow 窗口内最大对话次数
cooldownSeconds 相邻请求全局冷却
showThinking 是否展示思考过程
systemPrompt 系统提示词模板（{player_name} 占位符）

🚀 快速上手

/ai config set apiKey sk-xxxx
/ai config set webSearchEnabled true
/ai config set webSearchMode chatgpt
/ai 今天天气怎么样？

提示：联网功能是否生效取决于所连接模型/API 本身是否支持。DeepSeek 官方接口无原生联网，需通过兼容网关接入，或改用聚合平台预设（如 qwen）。

Building / 构建

Requires JDK 21.

./gradlew build

The built mod jar is at build/libs/severagent-<version>.jar.

License / 许可证

Licensed under the GNU General Public License v3.0 (GPL-3.0). See LICENSE.