liquipedia-dota2-mcp
Read-only MCP server for sourced Dota 2 tournament, team, player, schedule, and result data from Liquipedia, with revision-level metadata and caching.
README
liquipedia-dota2-mcp
Read-only MCP server and Codex skill for sourced Dota 2 tournament, team, player, schedule, and result data from Liquipedia.
This is an unofficial community project. It is not affiliated with or endorsed by Liquipedia, Team Liquid, Valve, or OpenAI.
Features
- Five structured MCP tools for entity search, tournament, team, player, and match queries.
- Revision-level source metadata, coverage indicators, and explicit parse warnings.
- Persistent SQLite cache, duplicate-request coalescing, and Liquipedia-compliant request queues.
- Local
stdiotransport for Codex, ChatGPT desktop, Claude Desktop, and other MCP clients. - Bundled
liquipedia-dota2Skill underskills/liquipedia-dota2.
Requirements
- Node.js 22 or newer. The server uses Node's built-in SQLite module and does not require a native dependency build.
- A custom
LIQUIPEDIA_USER_AGENTcontaining a project name, version, and contact URL or email.
Example:
MyDotaAssistant/1.0 (developer@example.com)
The server refuses to start without a valid identifying User-Agent. It only sends automated requests to https://liquipedia.net/dota2/api.php; it does not scrape rendered website pages.
Install and run
After the package is published to npm:
$env:LIQUIPEDIA_USER_AGENT = "MyDotaAssistant/1.0 (developer@example.com)"
npx -y liquipedia-dota2-mcp
Directly from GitHub, without npm publishing:
$env:LIQUIPEDIA_USER_AGENT = "MyDotaAssistant/1.0 (developer@example.com)"
npx -y github:Mooooooon/liquipedia-dota2-mcp
On npm 12 and newer, git package installs are disabled by default, so add --allow-git=root:
$env:LIQUIPEDIA_USER_AGENT = "MyDotaAssistant/1.0 (developer@example.com)"
npx -y --allow-git=root github:Mooooooon/liquipedia-dota2-mcp
dist/ is built automatically on install (prepare on npm 11 and earlier, pack-time prepack on npm 12), so the installer needs no build step.
From a source checkout:
npm install
npm run build
$env:LIQUIPEDIA_USER_AGENT = "MyDotaAssistant/1.0 (developer@example.com)"
node dist/cli.js
Logs use stderr so stdout remains a valid MCP transport.
Codex setup
Codex CLI and the ChatGPT desktop app share local MCP configuration. Add the published package with:
codex mcp add liquipedia-dota2 --env "LIQUIPEDIA_USER_AGENT=MyDotaAssistant/1.0 (developer@example.com)" -- npx -y liquipedia-dota2-mcp
Or add this to ~/.codex/config.toml:
[mcp_servers.liquipedia_dota2]
command = "npx"
args = ["-y", "liquipedia-dota2-mcp"]
startup_timeout_sec = 20
tool_timeout_sec = 120
[mcp_servers.liquipedia_dota2.env]
LIQUIPEDIA_USER_AGENT = "MyDotaAssistant/1.0 (developer@example.com)"
Install the bundled Skill by copying skills/liquipedia-dota2 into either the repository's .agents/skills directory or the user's .agents/skills directory. The Skill intentionally has no remote MCP dependency URL; configure the local server separately.
Claude Desktop setup
Add the server to the Claude Desktop MCP configuration:
{
"mcpServers": {
"liquipedia-dota2": {
"command": "npx",
"args": ["-y", "liquipedia-dota2-mcp"],
"env": {
"LIQUIPEDIA_USER_AGENT": "MyDotaAssistant/1.0 (developer@example.com)"
}
}
}
}
Tools
| Tool | Purpose |
|---|---|
search_dota2_entities |
Find tournament, team, and player candidates. |
get_dota2_tournament |
Read tournament metadata, participants, and placements. |
get_dota2_team |
Read team metadata, roster, and coaches. |
get_dota2_player |
Read player identity and team history. |
list_dota2_matches |
Read match schedule/result rows for a confirmed tournament or team. |
Every call returns data, section-level coverage, warnings, source, and a typed error. A section marked unavailable means the current parser could not establish the data; it does not mean that the source contains no records.
Cache and limits
The server enforces minimum intervals of 2.1 seconds for ordinary MediaWiki requests and 31 seconds for action=parse. These minimums cannot be weakened through environment variables. Defaults are:
- Search: 6 hours
- Team/player: 24 hours
- Upcoming or ongoing tournament: 30 minutes
- Completed tournament: 7 days
- Match list: 15 minutes
Set LIQUIPEDIA_CACHE_PATH to choose the SQLite file. When Liquipedia is unavailable, an expired cache entry may be returned with source.is_stale: true and its stale age.
Development
npm run check
npm test
npm run build
npm pack --dry-run
Default tests are completely offline. An opt-in live smoke test performs a small number of compliant API calls:
$env:LIQUIPEDIA_LIVE_TEST = "1"
$env:LIQUIPEDIA_USER_AGENT = "MyDotaAssistant/1.0 (developer@example.com)"
npm run test:live
Data and licensing
Project code is MIT licensed. Liquipedia text/data is separately provided under CC BY-SA 3.0 and remains subject to Liquipedia's API terms. Tool results include attribution, license, source page, revision, and fetch time. Media such as logos and player photographs is not downloaded or redistributed.
See Liquipedia API terms before deploying or modifying request behavior.
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。