Sequential orchestration error: Http 400 invalid request error, 'messages[0].name': string does not match pattern.

[bicheichane]

I'm getting an http 400 error when calling a simple sequential orchestration, right as i call GetValueAsync(), before any agents respond or my function filter gets hit

{"HTTP 400 (invalid_request_error: invalid_value)\r\nParameter: messages[0].name\r\n\r\nInvalid 'messages[0].name': string does not match pattern. Expected a string that matches the pattern '^[^\\s<|\\\\/>]+$'."}

This is with chat completion agents on azure open ai models, sequential orchestration with InProcessRuntime, everything updated to 1.60.

Passing in a query that looks pretty normal to me: "What are the three elements of deliberate practice that should be part of any effort to address ability, according to Ericsson's research?"

Any ideas what's going on or how to debug this?

[nicholasdbrady]

Hey @bicheichane! Based on your error message about "messages[0].name" not matching the required pattern when using Azure OpenAI chat completions, here's what could be causing the issue and some guidance on how to resolve it:

What’s Happening?

Your error:

HTTP 400 (invalid_request_error: invalid_value)
Parameter: messages[0].name

Invalid 'messages[0].name': string does not match pattern. Expected a string that matches the pattern '^[^\s<|\\/>]+$'.

This means that the name property on one of your chat completion message objects either:

• contains spaces, special characters like <, |, \, /, or >, or
• is empty or otherwise doesn't comply with the expected pattern.

Why This Matters

• The name property (if included) must be a valid string matching a narrow set of allowed characters—no spaces or special characters. For Azure OpenAI's chat completion endpoint, this is strict.
• If you are using system, user, or assistant roles, you generally don’t need a name field unless you are leveraging function calling.

Next Steps — Troubleshooting

1. Check Your Message Object:
   Make sure that none of your message objects look like this:

   {
     "role": "user",
     "content": "Hello",
     "name": "example name"  // <-- this is invalid, contains a space
   }

   Instead, if you include name, use only valid characters (e.g., "name": "assistant1").

2. Function Calling:
   If you are using function-calling, verify that the function name adheres to the restrictions above.

3. No name Needed?
   Try removing the name field entirely from your messages unless you absolutely need it. Most users do not need this property.

4. Other Special Tokens:
   Make sure your prompt/messages do not include legacy ChatML tokens or other special tokens, as these are now disallowed. See full details here:
   https://learn.microsoft.com/en-us/azure/ai-foundry/openai/how-to/chatgpt#troubleshooting

Official Microsoft Learn References

• Work with chat completions models > Troubleshooting

If you follow these steps and the error persists, try sharing the exact payload (with sensitive data removed) for more tailored help.

If this clears things up, please mark this as the answer! If you need more guidance, let me know!