This server offers a comprehensive, type-safe wrapper over the Telegram Bot API with useful engineering for AI use, but it is community-maintained and may lag behind the latest API versions.
Strongest · Setup experience 13/20 Weakest · Reliability 8/20
Static review · not runListed 2026-09-08
Read the FMRS scoring method →Runtime：Node.js
TELEGRAM_BOT_TOKEN
Required bot token, obtained from @BotFather after creating a Telegram bot.
TELEGRAM_DEFAULT_CHAT_ID
Optional default chat ID so tools don't need chat_id on every call.
TELEGRAM_DEFAULT_THREAD_ID
Optional default forum topic thread ID.
TELEGRAM_META_MODE
Set to true to expose only 2 meta-tools (telegram_find, telegram_call) instead of 169, saving ~99% context tokens; default false.
TELEGRAM_GLOBAL_RATE_LIMIT
Max requests per second, default 30, matching Telegram's global limit.
TELEGRAM_PER_CHAT_RATE_LIMIT
Max messages per minute per group, default 20, matching Telegram's limit.
TELEGRAM_MAX_RETRIES
Retry attempts on transient errors, default 3.
TELEGRAM_CB_THRESHOLD
Number of failures before the circuit breaker opens, default 5.
TELEGRAM_CB_COOLDOWN
Circuit breaker cooldown in milliseconds, default 30000.
TELEGRAM_ALLOWED_UPLOAD_DIRS
Comma-separated allowed upload paths for file upload security.
TELEGRAM_MAX_FILE_SIZE
Max upload file size in bytes, default 52428800 (50MB).
command: 'npx', args: ['telegram-api-mcp'], env: { TELEGRAM_BOT_TOKEN: 'your_token' }.npx telegram-api-mcp locally with an MCP client that supports stdio.{
"mcpServers": {
"telegram": {
"command": "npx",
"args": [
"telegram-api-mcp"
],
"env": {
"TELEGRAM_BOT_TOKEN": "your_token_from_botfather"
}
}
}
}
Shown for Claude Desktop. Other clients may use a different file or key (VS Code uses "servers") — the configurator below converts it.
{
"servers": {
"telegram": {
"command": "npx",
"args": [
"telegram-api-mcp"
],
"env": {
"TELEGRAM_BOT_TOKEN": "your_token_from_botfather"
}
}
}
}
Goes in your project's .vscode/mcp.json (VS Code uses a "servers" key).
claude mcp add telegram -e TELEGRAM_BOT_TOKEN=your_token_from_botfather -- npx telegram-api-mcp
Run it in a terminal; replace any <…> placeholders with your own values first.
After setup, the client's tool list should show tools like getMe and sendMessage; asking the AI to call getMe and getting back the bot's identity confirms the connection works.
Once connected, you can ask your AI assistant things like:
No matching tools
Listed from the project's documentation, not tested by this site.
telegram-api-mcp is a TypeScript-only MCP server that provides all 169 methods of the Telegram Bot API (Bot API v9.6; v10.2 per the manifest). It includes Zod parameter validation, token masking, tool annotations, response truncation, and resilience features like async-mutex token-bucket rate limiting (defaults: 30 requests/sec global and 20 messages/min per group, aligning with Telegram's official limits) and a 3-state circuit breaker (closed/open/half-open). It supports a meta-mode (TELEGRAM_META_MODE=true) that replaces the 169 individual tools with just two tools, telegram_find and telegram_call, to save ~99% of context tokens. It has only two runtime dependencies (@modelcontextprotocol/sdk and zod). The server communicates with clients via stdio. It is compatible with clients such as Claude Desktop, as per the README installation instructions, running through npx.
MCP server providing web search and content fetching via DuckDuckGo
Your AI never forgets again.
Telegram MCP server powered by Telethon to let MCP clients read chats, manage groups, and send/modify messages, media, contacts, and settings.
Official MCP server for Hostinger API.
Source revision 9d95026479c3 Data synced 2026-10-19 Read the FMRS scoring method