Why is the AI Agent in n8n is Extremely Slow When Processing Data?

Hello! I have set up a multi-agent system, where one agent has other agent nodes as tools.

It takes a long time, but especially a sub-agent that searches in an MCP. This MCP searches in a product catalog and the MCP is fast, but when it returns the response to the agent, it takes 30 seconds to return the result, which should be a JSON.

What could be happening? Should I set up the system differently?

Experience the same issue. Would be great to see an official fix for that.

Looking for an official fix as well.

Hi, facing the same problem (OpenAi/Gemini the same behaviour). Using self hosted/Docker/n8n Version 1.122.5. Did anyone try this with n8n V2?

Same here, I haven’t found a workaround in V2

Did you find a solution for cloud n8n?