Hey everyone! I’m working on creating an AI consultant in Telegram and got a bit stuck with workflow implementation.
My Current Task: Accumulate user messages for 10-15 seconds before sending to the AI model to:
- Gather full conversation context
- Reduce API call frequency
- Improve response quality
What I’ve Already Tried:
- Using Telegram Trigger
- Planning to use OpenAI/OpenRouter for response generation
- Want to implement temporary message storage in workflow
Questions for the Community:
- How to properly implement message buffering in N8N?
- What are the best practices for handling sequential messages?
- Are there any ready-made templates or examples for such a workflow?
I’ll be grateful for any advice and guidance! 
My Tech Stack:
- N8N
- Telegram
- OpenAI/OpenRouter