Frontend cannot parse messages from the llama.cpp chat UI

[NovaH00]

Here's the request made from the llama.cpp chat UI

{
  "messages": [
    {
      "role": "user",
      "content": "yo"
    },
    {
      "role": "assistant",
      "content": "Yo! What's up? How can I help you today? 😊"
    },
    {
      "role": "user",
      "content": "who are you?"
    }
  ],
  "stream": true,
  "return_progress": true,
  "reasoning_format": "auto",
  "backend_sampling": false,
  "timings_per_token": true
}

For some reasons, the tracing for this request doesn't work properly