VK GPT-4.1 Mini Reply Automation with n8n Workflow

Turn VK Messages Into Instant, GPT-4.1 Mini Replies (with n8n)

This n8n workflow automatically receives new VK messages via webhook, transcribes voice notes with OpenAI, generates a friendly assistant response using gpt-4.1-mini with per-user conversation memory, and sends the reply back to the same VK user.

What this workflow does

• Receives VK callback events through a POST webhook endpoint (path: /webhook/testvkwildbots) with message_new enabled.
• Prevents duplicate processing by checking the VK retry header and dropping retry deliveries when applicable.
• Routes the message based on content type:
  • If it contains text, it uses the original text as the input prompt.
  • If it includes a voice attachment, it downloads the MP3 from VK.
• Transcribes voice to text using OpenAI transcription.
• Generates an assistant reply using an OpenAI-powered agent (short-term conversation context per user).
• Sends the response back to VK via messages.send; if generation fails, it sends a fallback error message.

Use cases

• Answering VK user questions instantly in customer support or community management.
• Handling voice notes by converting them to text and replying in the same chat.
• Reducing manual workload for SaaS operators running VK-based onboarding or FAQs.

Technical details (n8n + integrations)

• VK webhook (POST callback API) with VK API version v=5.199 and messages.send endpoint.
• HTTP Request node to download VK MP3 attachments.
• OpenAI credentials for gpt-4.1-mini chat generation and audio transcription.
• Workflow nodes include Webhook, Set, Switch, Filter, HTTP Request, and n8n-nodes-langchainagent for agent-based replying.

Ideal if you’re building VK automation in n8n and want reliable, per-user conversational GPT replies.