n8n AI Telegram Bot: Smart Assistant & Content Summarizer

Turn Telegram messages into input, route commands to article summarization or image generation branches, and return results via Telegram

Who

Teams, content creators, and researchers who need quick article summaries or image prompts inside Telegram

You get

Deploy a 24/7 Telegram bot supporting /help, /summary <URL>, and /img <prompt> commands

You need

n8n · telegram

Steps

1. Prepare credentials and n8n environment

   Ensure you have access to an n8n instance (Cloud or self-hosted). Contact @BotFather in Telegram to create a bot and obtain the Bot Token. Get your OpenAI API key from the OpenAI platform.

2. Import the workflow template

   In n8n, open Templates, search for ID 4457 or title 'AI Telegram bot agent', then click Use this workflow to import it into your workspace.

3. Configure Telegram Trigger (Listener)

   Locate the Listener node (Telegram Trigger). Select or create Telegram credentials and enter the Bot Token. Confirm Updates includes 'message'. Saving automatically registers the webhook.

4. Configure Command Router

   Locate the Command Router node (If node). Check the condition: message.text startsWith '/help'. If true, route to Help Responder; if false, proceed to Summary Checker.

5. Configure Help Responder

   Locate the Help Responder node (Telegram node). Confirm it uses the same Telegram credentials. This node sends a Markdown-formatted help menu.

   Execution prompt · Source text。Help Responder node text parameter

   Copy prompt

   🤖 *Help Menu*
   
   Use `/summary <link>` to summarize an article.
   Use `/img <prompt>` to generate an image.
   
   _Example:_
   /summary https://example.com
   /img a futuristic cityscape

6. Configure summary check and fetch (Summary Checker → Fetcher → Text Extractor)

   Summary Checker (If node) checks if message.text startsWith '/summary'. The true branch goes to Fetcher (HTTP Request node), which retrieves webpage HTML from the Telegram message's link_preview_options.url with User-Agent: Mozilla/5.0. Then Text Extractor (HTML node) extracts body content using cssSelector=body and skipSelectors=svg,a.

7. Configure AI summarizer (Summarizer)

   Summarizer is an OpenAI node (typeVersion=1.8). You must manually select a model in modelId (currently empty). The messages field contains the system prompt with injected article text.

   Execution prompt · Source text。Summarizer node messages.values[0].content parameter

   Copy prompt

   =Summarize the entire content provided below into 10–12 concise bullet points. Ensure each point captures a unique and important aspect of the information, covering the core ideas, key facts, major findings, and essential takeaways. Avoid repetition and use clear, professional language suitable for quick understanding by a decision-maker.
   
   Content:
    {{ $json.text }}

8. Configure summary sender (Summary Sender)

   Summary Sender (Telegram node) returns the Summarizer output to the user. The text field references candidates[0].content.parts[0].text from Summarizer; chatId references Listener's message.chat.id.

   Execution prompt · Source text。Summary Sender node text parameter (expression)

   Copy prompt

   ={{$json["candidates"][0]["content"]["parts"][0]["text"]}}

9. Configure image command branch (Image Prompt Checker → Image Generator → Image Acknowledger)

   Image Prompt Checker (If node) checks if message.text startsWith '/img'. The true branch goes to Image Generator (OpenAI node, resource=image), which submits the prompt to an image generation API. Then Image Acknowledger (Telegram node) sends a confirmation message. Note: In this template, the Gemini image model does not return image URLs directly; actual image generation requires additional configuration with APIs like Stability.

   Execution prompt · Source text。Image Acknowledger node text parameter

   Copy prompt

   🖼️ Generated image prompt submitted! Gemini image model doesn't return images directly. Use image generation APIs like Stability for actual image URLs.

Build a Telegram AI bot in n8n using the following materials:
1. Prepare Telegram Bot Token (@BotFather) and OpenAI API key;
2. Import template 4457;
3. Configure Listener trigger, Command Router, and Help Responder;
4. Configure summary branch: Summary Checker → Fetcher (HTTP Request, url=$json.message.link_preview_options.url) → Text Extractor (HTML, body) → Summarizer (OpenAI, select model, enter prompt) → Summary Sender;
5. Configure image branch: Image Prompt Checker → Image Generator (OpenAI resource=image) → Image Acknowledger;
6. Save and activate the workflow, then test /help, /summary <URL>, and /img <prompt> in Telegram.