telebiz-mcp

What This Skill Does

The telebiz-mcp skill is a powerful bridge between the OpenClaw AI agent and your Telegram account. By utilizing the telebiz-tt browser client, this integration allows your AI agent to act as a personal assistant, managing your communications with precision. It enables the agent to list chats, retrieve message history, perform complex searches, manage folder categorization, and send messages directly through an authenticated Telegram session via a WebSocket relay. It essentially turns your Telegram client into a programmable, AI-driven workflow engine.

Installation

To get started, ensure you have Node.js 18+ installed. First, install the package globally using npm install -g @telebiz/telebiz-mcp. Next, log into your Telegram account at https://telebiz.io. Navigate to your local machine, navigate to your skill directory, and execute ./start-http.sh to initialize the persistent server on port 9718. Finally, enable 'Local MCP' within the Telebiz web settings under the Agent tab. You can verify your connection by running npm run health to confirm both the relay and browser executor are connected.

Use Cases

This skill is ideal for power users and professionals who want to automate their outreach or information management. You can use it to summarize long group chats, extract action items from direct messages, organize important conversations into specific folders automatically, or even set up automated reminders that get dispatched directly to your Telegram contacts. It is also useful for cross-referencing your CRM data with your active Telegram conversations.

Example Prompts

1. "List the 5 most recent unread messages from my 'Work' folder and summarize the key action items for each."
2. "Search for all messages from John Doe in the last 24 hours that contain the keyword 'invoice' and draft a polite reply asking for an ETA."
3. "Move all chats currently in my 'General' folder that are related to 'Project X' into a new 'Project X' folder."

Tips & Limitations

• Rate Limiting: Always respect the rate limits. You are restricted to 20 calls per request and 30 calls per minute. Ensure there is a 500ms delay between standard operations and 1s for heavy operations.
• Batching: The batchAddToFolder function contains a known bug. Avoid using it; instead, iterate through your chats and use addChatToFolder sequentially to ensure stability.
• CRM Stability: The linkEntityToChat feature is currently considered experimental and unreliable due to upstream schema inconsistencies; avoid mission-critical workflows involving this function until further notice.