How are you running AnythingLLM?

AnythingLLM desktop app

What happened?

I'm using AnythingLLM Desktop version on macOS and have encountered an issue where the thinking/reasoning content from the DeepSeek-R1:8b model is not displayed in the chat interface, even though it's clearly present in the raw API response from Ollama.

1. Setup Environment:

   • macOS Tahoe 26.0.1
   • AnythingLLM Desktop Version
   • Ollama Version 0.12.6
   • Model: deepseek-r1:8b

2. Configuration:

   • Ollama running on http://127.0.0.1:11434
   • DeepSeek-R1:8b model pulled and running correctly
   • AnythingLLM workspace configured to use this model

3. Observed Behavior:

   • In AnythingLLM chat: Only final response is displayed, thinking process is missing
   • In Ollama CLI or API: Both thinking process and final response are present

Evidence & Testing

Direct API Test (Working):

curl -X POST http://localhost:11434/api/generate -d '{
  "model": "deepseek-r1:8b",
  "prompt": "Please introduce yourself",
  "stream": false
}'

### Are there known steps to reproduce?

_No response_