Build an AI Task Manager in Telegram Using n8n and MCP Server

The Problem: Managing tasks across different platforms feels like juggling flaming torches while riding a unicycle. You're switching between your task manager app, checking emails, writing things down, and somehow still forgetting that important meeting. What if you could just talk to your assistant and have everything handled automatically?

The Solution: This n8n workflow transforms Telegram into your personal AI-powered task management assistant. Send a text message or voice note to your bot, and watch as it intelligently creates, updates, and manages tasks in Google Tasks. It's like having a smart assistant in your pocket, minus the coffee breaks.

Prerequisites: What You'll Need

Before diving into building this workflow, gather these essentials:

Required APIs and Accounts

• Telegram Bot Token: Create a bot using BotFather on Telegram to get your API credentials
• Google Tasks API: Enable Google Tasks API and set up OAuth2 authentication
• OpenAI API Key: For the GPT-4o-mini model that powers the AI agent
• Google Gemini API: Required for voice transcription capabilities
• n8n Instance: Either self-hosted or a cloud account to build the workflow

Key Components in This Workflow

This automation uses several powerful n8n nodes working in harmony:

• Telegram Trigger: Listens for incoming messages from your Telegram bot
• AI Agent: The brain of the operation, processing natural language requests
• MCP Server & Client: Connects the AI to your task management tools
• Google Tasks Nodes: Five different configurations for creating, updating, and retrieving tasks
• OpenAI Chat Model: Provides intelligent understanding of your requests
• Google Gemini: Transcribes voice messages into text
• Memory Buffer: Remembers your conversation context for natural interactions

Building Your AI Task Manager: Step-by-Step

Step 1: Set Up Message Reception and Voice Processing

The workflow starts with a Telegram Trigger node that monitors your bot for incoming messages. But here's where it gets interesting: not everyone wants to type out their tasks. The workflow includes a Switch node that detects whether you've sent a text message or a voice note.

When a voice message arrives, the workflow downloads the audio file through Telegram's API, then passes it to Google Gemini for transcription. The transcribed text gets formatted and sent forward, just like a regular text message. Your choice: type or talk. The system handles both seamlessly.

Pro Tip: Voice notes are perfect for capturing quick thoughts while you're on the go. The Gemini transcription is remarkably accurate, even with background noise.

Step 2: Configure Your AI Agent with Memory

The AI Agent node is where the magic happens. Powered by OpenAI's GPT-4o-mini model, it understands natural language and determines what action you want to take. The system message instructs the AI to check existing tasks before making updates, preventing duplicate entries and ensuring accuracy.

img_1.png

Connected to the agent is a Simple Memory node that maintains conversation context using a 20-message window. This means you can have natural back-and-forth conversations: "Add a task," "Actually, change the due date to Friday," "And add a note about calling the client first." The AI remembers what you're talking about.

The system message includes a crucial instruction: "Whenever asked to update a task, call the get_tasks tools first to retrieve the appropriate task ids then use that to update the tasks." This prevents the AI from guessing task IDs.

Step 3: Build Your MCP Server with Google Tasks Tools

The Model Context Protocol server acts as a bridge between your AI agent and Google Tasks. It exposes five specialized tools to the AI:

create_todays_task: Adds tasks with today's date automatically

create_upcoming_task: Creates tasks scheduled for future dates

complete_task: Marks tasks as done using their task ID

get_todays_tasks: Retrieves all tasks due today

get_upcoming_tasks: Fetches tasks scheduled for later

Each tool is a Google Tasks node configured with specific operations. The AI agent calls these tools based on your natural language requests. Say "Show me what I need to do today," and the AI knows to use the get_todays_tasks tool.

img_2.png

Step 4: Configure the MCP Client Connection

The MCP Client node connects your AI Agent to the MCP Server through a Server-Sent Events endpoint. This creates a real-time connection where the AI can discover available tools, understand their parameters, and execute them on demand.

The setup uses a unique SSE endpoint specific to your n8n instance. Once connected, the AI automatically learns about all five Google Tasks tools without you having to explain each one individually. It's like introducing your AI to a toolbox and letting it figure out which tool to use for each job.

Step 5: Handle Responses and Send Messages Back

After the AI Agent processes your request and executes the necessary tools, it generates a human-friendly response. The chatOutput node extracts this response and formats it properly. Finally, the sendMessage node delivers the reply back to your Telegram chat.

img_3.png

The workflow ensures attribution is disabled to keep responses clean and sets notification preferences so you're not bombarded with alerts. The entire process, from receiving your message to sending a response, typically takes just a few seconds.

img_4.png

Real-World Benefits and Use Cases

Why This Workflow Changes Everything

Traditional task managers require context switching. You stop what you're doing, open an app, navigate to the right list, fill out forms, and return to your work. This workflow eliminates that friction. Telegram is already open on most people's phones and computers, making task management as simple as sending a message to a friend.

Practical Applications

• Meeting Notes to Tasks: During a meeting, quickly voice-message your bot with action items without breaking focus
• Email Overload: Forward important emails to yourself via Telegram, then tell your bot to create tasks from them
• Morning Planning: Ask "What do I have today?" and get a complete overview without opening multiple apps
• Team Coordination: Share your bot with team members for collaborative task management in a familiar interface
• Evening Reviews: Check off completed tasks by simply messaging "Mark meeting prep as done"

Scalability and Customization

The beauty of this n8n workflow lies in its flexibility. Want to add priority levels? Create another tool in the MCP server. Need integration with Notion instead of Google Tasks? Swap out the nodes. Want to add calendar integration? Connect a Google Calendar trigger. The architecture supports endless expansion.

The workflow runs continuously in the background, always ready to respond. With proper error handling and the reliability of n8n's execution engine, you get a production-ready task management system that scales with your needs.

Your AI Assistant Awaits

Building an AI-powered task manager with n8n transforms how you interact with productivity tools. By combining Telegram's ubiquity, Google Tasks' reliability, and OpenAI's intelligence, you create something greater than the sum of its parts: a truly conversational interface for managing your work.

The workflow handles both text and voice inputs, remembers your conversation context, and intelligently routes requests to the right tools. Whether you're creating tasks for today, scheduling future work, or reviewing what needs attention, your AI assistant understands and responds appropriately.

Start with this foundation and customize it to match your workflow. Add more tools, integrate different services, or enhance the AI's capabilities. The hard part is already done. Now you just need to send that first message and watch your personal AI assistant spring to life.