Vexa
Meeting bot and transcripts for Google Meet, Teams and Zoom. Live or after, speakers labelled.
作者提供的说明(英文原文)
MCP server for Telegram serverless bots: lifecycle, deploys, migrations, webhooks via tgcloud CLI
这三项永远不会合并成一个分数:意见、测量和人气是不同的东西。
我们还没有为它写通俗的介绍。下面是我们从注册中心读取到的事实。
选择你使用的 AI 应用,然后复制设置。
TGCLOUD_CLI, TGCLOUD_CLI_ARGS, TGCLOUD_TIMEOUT_MS, TGCLOUD_TOKEN标为机密的值是你账户的钥匙。千万不要分享给别人,也不要粘贴到聊天里。
你需要
你用的是哪个应用?
在上方选择你的应用。我们会记住它,下次按钮会显示“安装到 …”。
打开 Claude Desktop → Settings → Developer → Edit Config,把下面的内容粘贴到 claude_desktop_config.json
{
"mcpServers": {
"tgcloud": {
"command": "npx",
"args": [
"-y",
"tgcloud-mcp"
],
"env": {
"TGCLOUD_CLI": "your-value",
"TGCLOUD_CLI_ARGS": "your-value",
"TGCLOUD_TIMEOUT_MS": "your-value",
"TGCLOUD_TOKEN": "your-value"
}
}
}
}在终端中运行这条命令
claude mcp add tgcloud --env TGCLOUD_CLI=your-value --env TGCLOUD_CLI_ARGS=your-value --env TGCLOUD_TIMEOUT_MS=your-value --env TGCLOUD_TOKEN=your-value -- npx -y tgcloud-mcp运行前请先填写空着的值。
粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json
{
"mcpServers": {
"tgcloud": {
"command": "npx",
"args": [
"-y",
"tgcloud-mcp"
],
"env": {
"TGCLOUD_CLI": "your-value",
"TGCLOUD_CLI_ARGS": "your-value",
"TGCLOUD_TIMEOUT_MS": "your-value",
"TGCLOUD_TOKEN": "your-value"
}
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
粘贴到项目中的 .vscode/mcp.json
{
"servers": {
"tgcloud": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"tgcloud-mcp"
],
"env": {
"TGCLOUD_CLI": "your-value",
"TGCLOUD_CLI_ARGS": "your-value",
"TGCLOUD_TIMEOUT_MS": "your-value",
"TGCLOUD_TOKEN": "your-value"
}
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
这个服务器在你的电脑上运行,而 ChatGPT 只能连接互联网上的服务器。请选择其他应用。
你的选择通过 Cookie 保存在这台设备上,无需账户。
TGCLOUD_CLI envTGCLOUD_CLI_ARGS envTGCLOUD_TIMEOUT_MS env这些设置是根据作者在注册中心的条目生成的。如果 README 的写法不同,请以 README 为准。
它在你的电脑上运行;安装后,工具列表会显示在你的 AI 应用中。
我们检查的是任何人都能看到的信息;我们没有阅读代码。不适用的检查项不计入总数。 · 检查于 2026年11月8日
注册名称已通过 GitHub 账户 sdamarketing 验证,该账户拥有此仓库
我们还没有从 GitHub 读取这个仓库
我们还没有从 GitHub 读取这个仓库
它在你的电脑上运行,没有互联网端点
在你的电脑上运行,改用本地设置
它在你的电脑上运行;我们只能读取互联网端点的工具列表
与注册中心里这个分类中所有活跃的服务器比较。只有当发布者拥有该仓库时,星标数才计入。
它在 1 项适用的基础检查中通过了 1 项。我们没有阅读代码。在添加密钥或令牌之前,请先看看这个服务器要求哪些权限,并使用只具备所需权限的密钥。
你需要填写:TGCLOUD_TOKEN
它在你的电脑上运行。你的 AI 应用会在需要时启动它,数据不会经过作者的服务器(除非它自己调用在线服务)。
GitHub 仓库最后更新于–,最新版本(0.3.3)于23天前进入注册中心。
支持 MCP 的应用,例如 Claude Desktop、Claude Code、Cursor 和 VS Code。你的电脑上还需要有 Node.js, Docker。
同一分类中有:Vexa, emailmd, MCP WhatsApp Cloud
同一分类中其他作者的服务器,按星标数排序。
如何看卡片:★ GitHub 星标i三大支柱:评价 · 检查 · 使用情况i更新 = 最新版本发布到注册中心的时间i
Meeting bot and transcripts for Google Meet, Teams and Zoom. Live or after, speakers labelled.
Render markdown into email-safe HTML, lint drafts for deliverability problems, and preview emails.
MCP server for WhatsApp Cloud API (Meta) — messages, media, and templates via Graph API
按原文从作者的仓库复制。PopNx MCP 没有审查或背书其内容。HTML、图片和脚本已被移除。链接会离开本站。
Русский · 中文 · العربية · Deutsch · Français
Unofficial project. This is a community tool (author @sdamarketing),
not affiliated with Telegram. The platform's official tool is the
@tgcloud/cli
CLI. tgcloud-mcp wraps that CLI in MCP. Threat model:
SECURITY.md.
Teaches an AI assistant to manage Telegram serverless bots (core.telegram.org/bots/serverless): you say "create a bot", "deploy", "show the webhook" — the assistant runs tgcloud and you do not have to drive the terminal.
Works with any MCP client: opencode, Claude Code, Claude Desktop, Cursor, VS Code (Copilot), Windsurf, Zed. The same server and skill are also one Agent Plugin (agent-plugin/) for Codex, Cursor, Copilot, Kiro, and VS Code.
The docs site (docs7) and the Context7 index: context7.com/sdamarketing/tgcloud_mcp.
| Section | Contents |
|---|---|
| Getting started | Overview, quickstart, installation |
| Usage | Tools, resources, security |
| Help | Troubleshooting, cookbook |
| Reference | SETUP, evening walkthrough |
Other languages: ru, zh, ar, de, fr. Commands: tgcloud-mcp setup, tgcloud-mcp update, tgcloud-mcp links, tgcloud-mcp --help.
16 tools over the tgcloud CLI:
| Area | Tools | What they do |
|---|---|---|
| Lifecycle | create_project, init_project, login_bot, add_module, upgrade | Scaffold, link a bot with the @BotFather token, new handlers, endpoints, and lib modules, move a pre-0.2.0 layout into tgcloud/ |
| Sync | status, diff, push, deploy, pull, fetch, reset | See changes, atomic deploy, build-then-push for a Mini App, sync with the cloud |
| Data | run_handler, migrate | Run a handler or endpoint without deploying; migrations with --dry-run, --safe, --yes, --local |
| Webhook | webhook_status, webhook_sync | "The bot is silent", re-point the webhook |
Plus MCP resources tgcloud://docs/* — platform reference in English (layout, import rules, sdk/db, sdk/api, sdk/fetch, CLI, limits).
Tool descriptions and the skill are English so the model follows one text. The assistant is told to answer you in your language. Human docs are English, Russian, Chinese, Arabic, German, and French.
reset, push --force, applying upgrade, applying migrations, and webhook sync --drop-pending refuse to run without confirm: true. The assistant shows the change and asks first.Prefer the npm package (published with OIDC trusted publishing and Sigstore provenance). curl | bash is convenient and runs a remote script: read it first (drop the | bash), or use npm.
Node.js 20+ (the platform CLI needs 18+).
npm:
npm install -g tgcloud-mcp
tgcloud-mcp setup
The wizard writes the MCP config. Say yes to the Agent Plugin and it copies agent-plugin/ to ~/.tgcloud-mcp/agent-plugin, into Cursor's local plugin folder, and into VS Code chat.pluginLocations.
One command (macOS / Linux / WSL):
curl -fsSL https://raw.githubusercontent.com/sdamarketing/tgcloud_mcp/main/install.sh | bash
Docker:
docker run -i --rm ghcr.io/sdamarketing/tgcloud-mcp
From source:
git clone https://github.com/sdamarketing/tgcloud_mcp.git
cd tgcloud_mcp npm install && npm test
Details: [docs/SETUP.md](docs/SETUP.md).
{
"mcpServers": { "tgcloud": { "command": "tgcloud-mcp" } } }
| Variable | Default | Purpose |
|---|---|---|
TGCLOUD_CLI | npx | How to launch the CLI (a global tgcloud is fine) |
TGCLOUD_CLI_ARGS | tgcloud | Arguments before the subcommand |
TGCLOUD_TIMEOUT_MS | 120000 | Timeout of one CLI command |
You: create an echo bot in ~/bots/echo
Agent:
create_project→login_bot(asks @BotFather for the token) → editshandlers/message.js→push→webhook_sync. The bot is live.
You: test it
Agent:
run_handlerwith{ chat: { id: 1 }, text: "hi" }and shows the output and logs.
After front-end edits the agent calls deploy (build, then push). A plain push uploads the build folder as it is and does not run the bundler.
[A serverless bot with a Mini App in one evening](docs/cookbook-srl-bot.md) — the long walkthrough is Russian; the commands are the platform's. Short overviews: docs/cookbook.en.mdx and the ru, zh, ar, de, fr siblings.
skills/tgcloud is the procedure the agent follows: layout, imports, recipes, guardrails. The same file is agent-plugin/skills/tgcloud/SKILL.md. npm run docs:check fails if the copies diverge.
npx skills add sdamarketing/tgcloud_mcp
src/ # server, runner, tools
resources/docs/ # platform reference → MCP resources (English) skills/tgcloud/ # agent skill agent-plugin/ # Agent Plugins 1.0 package (plugin.json, mcp.json, skill) docs/ # human docs, six languages
MIT