AI-Powered Telegram Agent: Intelligent Text, Audio & Image Processor

Unlock Intelligent Telegram Interactions with this AI Agent

This n8n AI Agent connects to your Telegram bot via webhook to receive various message types (text, voice, images). It then leverages OpenAI's powerful models to understand and process these messages. Audio messages are transcribed to text, images are analyzed for their content, and text messages (including transcribed audio) are classified by intent (e.g., 'task' or 'other') using AI. Based on the message type and AI processing, the agent can then trigger different follow-up actions or replies directly within Telegram.

Key Features & Benefits

• Multimodal Message Handling: Seamlessly ingests and processes text, voice notes, and images sent to your Telegram bot.
• AI-Powered Audio Transcription: Utilizes OpenAI (Whisper model via Langchain node) to accurately convert voice messages into text, making audio content actionable.
• Intelligent Image Analysis: Leverages OpenAI's vision capabilities (e.g., gpt-4o-mini) to understand and generate descriptions of image content.
• AI Text Classification: Employs an AI-driven text classifier (Langchain with an OpenAI LLM like gpt-4o-mini) to categorize messages by intent, enabling smarter routing and responses.
• Automated Routing & Response: Directs processed information down different paths for customized replies or further actions within Telegram or other integrated systems via n8n.
• Optional User Validation: Includes a pre-built step to verify the sender's identity based on Telegram user details for more secure or personalized interactions.
• Webhook Management Utilities: Provides helper nodes and guidance for setting up, testing, and checking your Telegram bot's webhook status.

• An n8n instance (Cloud or self-hosted).
• OpenAI API Key with access to models like gpt-4o-mini (or other suitable models for transcription, image analysis, and chat/classification).
• Telegram Bot Token (obtainable from BotFather by creating a new bot).
• A publicly accessible URL for your n8n webhook endpoint (e.g., https://[your-n8n-instance-url]/webhook/[your-chosen-path]). This is where Telegram will send message updates.

1. Download the n8n workflow JSON file.
2. Import the workflow into your n8n instance.
3. Configure the 'Telegram Token & Webhooks' (Set node):
   • Replace bot[your-telegram-bot-token] with your actual Telegram Bot Token.
   • Update test_url and production_url with your n8n instance's public webhook URL. The path (e.g., /webhook/your-endpoint) should match the path configured in the 'Listen for Telegram Events' (Webhook node).
4. To register your webhook with Telegram:
   • Temporarily connect the 'Telegram Token & Webhooks' node to either 'Set Webhook Test URL' or 'Set Webhook Production URL' and run it once.
   • Alternatively, use the 'Get Telegram Webhook Info' node (after setting the token) and follow the manual setup instructions found in the large sticky note if preferred.
5. In the 'Validation' (Set node):
   • Enter the first_name, last_name, and id of specific Telegram user(s) if you want to restrict bot interactions to them.
   • If no user validation is needed, you can disable or modify the 'Check User & Chat ID' (If node).
6. Ensure your OpenAI API credentials are correctly configured in all OpenAI and Langchain nodes: 'gpt-4o-mini', 'gpt-4o-mini1' (for Text Classifiers), 'Analyze Image', and 'Transcribe Recording'.
7. Customize the categories and descriptions in the 'Text Classifier Audio' and 'Text Classifier' nodes to match your specific use case (e.g., support, sales, general query).
8. Review and adjust the content and logic of the final Telegram reply nodes (e.g., 'Audio Task Message', 'Text Other Message', 'Image Message') to suit your desired bot responses.
9. Activate the workflow. Send a text, voice message, or image to your Telegram bot to test.