Simple N8N agent always response "Error: Failed to receive response"

[skiligge]

This is making me sick for hours and hours, I can’t even get an answer. I am trying to test n8n. I have credits in anthropic and openapi, but on the LLMs of both it does not give me this result. It’s a fresh instance of n8n on Neubox.

[Anshul_Namdev]

Hi @skiligge Welcome!
For the LLMs try revoking your credentials and with fresh credentials try to access those models and then try to use a different API key with different account. Let me know if that works.

[Benjamin_Behrens]

Hey skiligge, “Failed to receive response” in n8n almost always means the instance can’t reach the OpenAI API at all — it’s a network issue, not an API key problem.

First test: try the HTTP Request node in n8n and call https://api.openai.com/v1/models with your Bearer token as the auth header. If that also fails, the problem is at network level.

Since you’re on Neubox, check if they restrict outbound HTTPS to third-party APIs on their plan — some hosting providers block this on lower tiers. You can also open the n8n execution log to see the raw error message, which often gives more detail than the chat widget error.

If the HTTP test succeeds but the agent still fails, double-check that you’re using the right OpenAI node version for your n8n release (Settings → About shows your n8n version).

What does the execution log say exactly?

[Brighto]

Works good in my side with the JSON you have provided.

[Benjamin_Behrens]

Thanks @Brighto for testing that — confirms the workflow itself is fine.

@skiligge this narrows it down: the JSON is correct, so the issue is 100% on the environment side. Either the Neubox instance can’t reach the OpenAI API, or there’s a credential issue specific to your setup.

Did the HTTP Request test I mentioned show anything? That would tell us immediately if it’s a network block.