I cant make a custom gpt connect to n8n trough an action. It just stopped working. Anyone that could help with that?
Hi @Lone_Brooks
Hard to tell what’s wrong without more detail, but a few common things to check:
→ In Credentials → OpenAI, hit Test and make sure your API key is still valid (401/Unauthorized usually means re-auth).
→ Confirm how you’re calling the custom GPT: OpenAI node, Chat Model/AI Agent, or HTTP Request. n8n recently updated the OpenAI node, so older “assistant” setups can stop working.
→ Some models reject certain parameters, so as a sanity check you can try a simple HTTP Request with the same key.
If you can share the exact error message, which node you’re using, and your n8n version, we can help you pinpoint the issue quickly.