Missing Models in OpenRouter Credentials on Latest n8n Cloud Version

Below is some guidance on troubleshooting this issue:

Check the Base URL Configuration

When models show up on your self-hosted instance but not via credentials on Ambassador Cloud, it’s a sign that the URL used by your credentials may not be returning the model list as expected. In some cases, ensuring that the correct base URL is configured in the credentials (or switching to specifying it directly on the node) can help resolve connectivity or retrieval issues.

Compare How the Instance Is Accessed

Since your self-hosted instance works when you override the base URL on the node, verify that your Ambassador Cloud configuration uses the equivalent endpoint. The documentation points out that using credentials to automatically load models requires that the endpoint in the credentials is reachable and returns the proper results. A mismatch in addressing (for example, due to network routing or binding issues) can result in an empty models list.

Verify Your Connection and API Key

1. Double-check that your OpenRouter API key and related settings in your credentials are correct.
2. Make sure that your Ambassador Cloud account hasn’t introduced changes (or rate limits) that might affect how the response is delivered.
3. If possible, try testing your OpenRouter endpoint with an HTTP Request node to see if it returns the expected list of models.

Consider Caching or Intermittent Issues

If it worked once and then stopped, it might be an intermittent network issue or a caching problem. Checking your n8n logs for errors or network timeouts can provide clues. Sometimes republishing ports or ensuring proper network configuration (especially when different hosting environments are involved) resolves such transient issues.

Summary

• Ensure your credentials (specifically the base URL) for Ambassador Cloud match what your self-hosted instance uses.
• Validate that your API key and account settings for OpenRouter are correct and haven’t triggered rate limiting or quota issues.
• Test connectivity using an HTTP Request node to see if the models endpoint is working as expected.

If after verifying these points the issue persists, you might be encountering an environment-specific problem with how the Ambassador Cloud instance handles model fetching. In that case, reaching out to the support team with these details can help pinpoint and resolve the problem.

Hope this helps you narrow down the issue!