teamspeak3-mcp
Let AI models manage your TeamSpeak 3 server through the Model Context Protocol.
README
<div align="center">
TeamSpeak 3 MCP Server
Let AI models manage your TeamSpeak 3 server through the Model Context Protocol.
</div>
A Model Context Protocol (MCP) server that exposes TeamSpeak 3 ServerQuery operations as AI-callable tools. Connect Claude, Cursor, or any MCP-compatible client to manage your TeamSpeak server with natural language.
Features
- 35 purpose-built tools covering server management, channels, clients, groups, permissions, moderation, and more
- Lazy connection — connects to TeamSpeak only when the first tool is invoked
- Exponential backoff retry — automatic reconnection with up to 3 attempts
- Graceful shutdown — cleans up the ServerQuery session on process exit
- Centralized error handling — every tool returns structured MCP error responses
- Zero-config transport — runs over
stdio, works out-of-the-box with any MCP client
Requirements
- Node.js >= 18
- A TeamSpeak 3 server with ServerQuery access (port
10011by default)
Getting Started
Add the following to your MCP client configuration. This works with most clients:
{
"mcpServers": {
"teamspeak": {
"command": "npx",
"args": ["teamspeak3-mcp"],
"env": {
"TEAMSPEAK_HOST": "your-server.com",
"TEAMSPEAK_PASSWORD": "your-password"
}
}
}
}
No installation needed — npx downloads and runs the package automatically.
Claude Desktop
Add to your Claude Desktop config file (claude_desktop_config.json):
{
"mcpServers": {
"teamspeak": {
"command": "npx",
"args": ["teamspeak3-mcp"],
"env": {
"TEAMSPEAK_HOST": "your-server.com",
"TEAMSPEAK_PASSWORD": "your-password"
}
}
}
}
Cursor
Add to your Cursor MCP settings (.cursor/mcp.json):
{
"mcpServers": {
"teamspeak": {
"command": "npx",
"args": ["teamspeak3-mcp"],
"env": {
"TEAMSPEAK_HOST": "your-server.com",
"TEAMSPEAK_PASSWORD": "your-password"
}
}
}
}
Configuration
Configuration is resolved from CLI arguments first, then environment variables, with sensible defaults as fallback.
| Parameter | CLI Flag | Env Variable | Default |
|---|---|---|---|
| Host | --host |
TEAMSPEAK_HOST |
localhost |
| Query Port | --port |
TEAMSPEAK_PORT |
10011 |
| Username | --user |
TEAMSPEAK_USER |
serveradmin |
| Password | --password |
TEAMSPEAK_PASSWORD |
(required) |
| Virtual Server ID | --server-id |
TEAMSPEAK_SERVER_ID |
1 |
| Enabled Tools | --tools |
TEAMSPEAK_TOOLS |
(all) |
Note:
TEAMSPEAK_PASSWORDrefers to the ServerQuery login password, not the TeamSpeak server connection password. You can find it in the server console output on first start, or create one via Tools → ServerQuery Login in the TeamSpeak client.
Selective Tool Loading
By default all 35 tools are registered. Use TEAMSPEAK_TOOLS (or --tools) with a comma-separated list of module names to load only what you need — useful for reducing the tool list exposed to the AI model:
{
"mcpServers": {
"teamspeak": {
"command": "npx",
"args": ["teamspeak3-mcp"],
"env": {
"TEAMSPEAK_HOST": "your-server.com",
"TEAMSPEAK_PASSWORD": "your-password",
"TEAMSPEAK_TOOLS": "server,channel,client"
}
}
}
}
Available modules: server, channel, client, sgroup, cgroup, permission, messaging, moderation, token, file
Tools Reference
Server (server_*)
| Tool | Description |
|---|---|
server_info |
Get server details (scope: overview or connection statistics) |
server_list |
List resources (resource: clients, channels, server_groups, channel_groups) |
server_search |
Search for clients or channels by pattern |
server_log |
View recent virtual server or instance log entries |
server_diagnose |
Run a diagnostic check on the current connection's permissions |
Channel (channel_*)
| Tool | Description |
|---|---|
channel_create |
Create a new channel (permanent or temporary) |
channel_delete |
Delete a channel (with optional force flag) |
channel_update |
Update channel properties (name, password, codec, talk power, etc.) |
channel_info |
Get detailed channel information |
channel_perm |
Add, remove, or list permissions on a channel |
Client (client_*)
| Tool | Description |
|---|---|
client_info |
Get detailed info: platform, version, country, IP, idle time, etc. |
client_move |
Move a client to another channel |
client_kick |
Kick a client from the server or channel |
client_ban |
Ban a client (timed or permanent) |
client_perm |
Manage server group membership and individual permissions |
client_db_list |
List historical clients from the server database (includes offline clients) |
client_poke |
Send a poke alert notification to a client |
Server Group (sgroup_*)
| Tool | Description |
|---|---|
sgroup_create |
Create a new server group |
sgroup_delete |
Delete a server group |
sgroup_perm |
Add, remove, or list permissions on a server group |
sgroup_clients |
List all clients assigned to a server group |
Channel Group (cgroup_*)
| Tool | Description |
|---|---|
cgroup_create |
Create a new channel group |
cgroup_perm |
Add, remove, or list permissions on a channel group |
cgroup_assign |
Assign a client to a channel group in a specific channel |
Permission (perm_*)
| Tool | Description |
|---|---|
perm_list |
List all available permission definitions (name, ID, description) |
perm_find |
Find all assignments of a permission across the server |
perm_overview |
Get effective permission overview for a client in a channel |
Messaging (msg_*)
| Tool | Description |
|---|---|
msg_send |
Send a text message (mode: channel or private) |
Moderation (ban_* / complaint_*)
| Tool | Description |
|---|---|
ban_list |
List all active ban rules |
ban_manage |
Create, delete, or clear ban rules by IP/name/UID |
complaint_list |
List complaints (optionally filtered by target client) |
Tokens (token_*)
| Tool | Description |
|---|---|
token_list |
List all available privilege keys/tokens |
token_create |
Create a server group or channel group token |
Files (file_*)
| Tool | Description |
|---|---|
file_list |
List files in a channel's file repository |
file_info |
Get detailed info about a specific file |
Development
git clone https://github.com/fl0w1nd/teamspeak3-mcp.git
cd teamspeak3-mcp
pnpm install
pnpm build # Build the project
pnpm dev # Watch mode (auto-rebuild on changes)
pnpm inspect # Debug with MCP Inspector (web UI)
MCP Inspector
The project includes a pre-configured script for the MCP Inspector, a web-based debugging tool:
pnpm inspect
This launches a local web UI where you can browse available tools, invoke them interactively, and inspect request/response payloads — useful for development and troubleshooting.
Project Structure
src/
├── index.ts # Entry point, stdio transport, graceful shutdown
├── config.ts # CLI + env configuration parsing
├── connection.ts # TeamSpeak connection with retry & lazy init
├── server.ts # MCP server setup & tool registration
├── utils/
│ └── tool-handler.ts # Error handling & response utilities
└── tools/
├── server.ts # Server info, resource listing, search, logs, diagnostics
├── channel.ts # Channel CRUD & permissions
├── client.ts # Client management, permissions & poke
├── server-group.ts # Server group CRUD, permissions & members
├── channel-group.ts # Channel group CRUD, permissions & assignment
├── permission.ts # Global permission queries & overview
├── messaging.ts # Channel & private messaging
├── moderation.ts # Bans & complaints
├── token.ts # Privilege token management
└── file.ts # Channel file browser
License
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。