Agnes AI Chat, Image & Video Reply Assistant for Telegram (n8n)

Turn Telegram prompts into AI replies, images, or videos—automatically

Agnes AI Chat, Image & Video Reply Assistant for Telegram (n8n) is an n8n automation workflow that receives a chat message, routes it to Agnes AI, and then sends back the result to Telegram—either a text response, an uploaded image, or a finished video with a caption.

What this workflow does

Designed for prompt-based routing, the workflow performs these steps:

• Receives an incoming chat message and reads the user’s prompt.
• Routes the prompt by keywords to one of three paths: text chat completion, image generation, or video creation in Agnes AI.
• Text prompts: calls the Agnes AI chat completions endpoint and returns the generated reply to the chat.
• Image prompts: calls the Agnes AI image generation endpoint, then uploads the resulting image URL to ImageKit.
• Video prompts: starts a video generation job in Agnes AI, polls job status until completed, downloads the final video, and sends it to a specified Telegram chat as a video message with a caption.

Use cases

• Let users message your Telegram bot and receive instant AI text replies or automatically generated creative images.
• Generate video responses from prompts and deliver them directly inside a Telegram community or support channel.
• Create a branded content pipeline where images are stored via ImageKit and videos are delivered through Telegram.

Technical details

• Keyword routing using nodes such as if, switch, and wait
• Telegram node for sending video messages
• HTTP Request nodes to call Agnes AI endpoints with Agnes AI HTTP Bearer Auth
• ImageKit upload using provided upload credentials (replace the Authorization header value for the upload request)
• Workflow includes n8n-nodes-langchainchat for chat integration logic

Setup highlights

• Add Agnes AI HTTP Bearer Auth and ensure it’s selected for Agnes AI HTTP requests.
• Add a Telegram API credential and update the target Telegram chat ID for video delivery.
• Adjust the keyword routing rules (text/image/video) to match your prompt classification needs.