Cripto Norte Telegram MCP Server
Enables publishing posts to a Telegram channel via an MCP tool, designed for scheduled content delivery and integration with Claude.
README
Cripto Norte — Telegram MCP server
Remote MCP server with one tool, post_to_telegram_channel, that lets Claude
publish a finished post to your Telegram channel. Tested locally end-to-end
(MCP handshake, tool listing, tool call, error handling) — ready to deploy.
1. Create the Telegram bot
- In Telegram, open @BotFather →
/newbot→ follow the prompts. - Copy the token it gives you (
TELEGRAM_BOT_TOKEN). - Add the bot to your channel as an admin with permission to post messages.
- Your
TELEGRAM_CHAT_IDis your channel's@username(if public) or its numeric ID (if private — get it by forwarding a channel message to @userinfobot or similar).
2. Deploy the server
Any Node.js host that gives you a public HTTPS URL works. Easiest options:
Railway / Render (free tier works for this)
- Push this folder to a new GitHub repo.
- Create a new Web Service on Railway or Render, point it at the repo.
- Set the environment variables from
.env.example(real values) in the host's dashboard. - Build command:
npm install. Start command:npm start. - Once deployed, note the public URL, e.g.
https://your-app.up.railway.app. The MCP endpoint ishttps://your-app.up.railway.app/mcp.
Fly.io / a VPS work too — same env vars, same npm start, just expose
port 3000 (or whatever PORT you set) over HTTPS.
Locally, npm install && npm start runs it on http://localhost:3000/mcp for testing.
3. Connect it to Claude
- In Claude: Customize → Connectors → + → Add custom connector.
- Name:
Cripto Norte Telegram. - URL:
https://your-app.up.railway.app/mcp. - Advanced settings: since this server checks a simple Bearer token
(
MCP_AUTH_TOKEN) rather than full OAuth, check whatever your Claude plan's custom-connector UI currently offers for a static bearer token — this has changed recently, so if you don't see that option, it's worth checking support.claude.com for the current field name, or dropping theMCP_AUTH_TOKENcheck entirely for a first personal test. - Click Add, then Connect.
- In a chat, click + → Connectors and enable it for that conversation.
4. Test it
Ask Claude in that conversation: "send a test message to the Telegram
channel using the connector." It should call post_to_telegram_channel and
you'll see the message land in your channel.
5. Wire it into a scheduled task
In Claude Cowork → Scheduled → New task → Set up manually:
- Enable the
Cripto Norte Telegramconnector for the task. - Prompt: reference the
cripto-norte-postsskill, the post type (UTRNNIK / digest), where to pull current prices/news from, and an explicit instruction to callpost_to_telegram_channelwith the finished, formatted text. - Set the cadence (e.g. daily 08:30 and 19:00 local time).
- Start with approval mode "requires confirmation," switch to automatic once you've checked a few runs.
Notes
- The server keeps sessions in memory. If it restarts, Claude just re-initializes on the next call — no action needed.
parse_mode: "Markdown"is used when sending, matching the skill's formatting rules (monospace titles, bold figures, em dash in headers only).- If you ever need more tools (e.g. posting to multiple channels, or
separate tools per coin), add more
server.registerTool(...)blocks inindex.jsfollowing the same pattern.
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。