How to set up .env.local for deepseek-r1:32b and ollama

[iChristGit]

I am trying very hard to get it to perform a good web-search, but it cannot craft the querry with my current .env.local
I also tried other .env.local combination, but the assistant always respond like that and wont generate any good queries:

Can someone share their working ollama deepseek env file?

MONGODB_URL=mongodb://localhost:27017
USE_LOCAL_WEBSEARCH=true
PUBLIC_APP_ASSETS=chatui
HF_ACCESS_TOKEN=hf_none
PUBLIC_APP_DESCRIPTION="ChatGPT But Open Source!"
PUBLIC_APP_NAME=ChatGPT
WEBSEARCH_JAVASCRIPT=true
HF_TOKEN=none
MODELS=`[
  {
    "name": "DeepSeek-R1-32B",
    "parameters": {
      "temperature": 0.1,
      "top_p": 0.95,
      "repetition_penalty": 1.2,
      "top_k": 50,
      "truncate": 3072,
      "max_new_tokens": 1024,
      "stop": ["</s>"]
    },
    "endpoints": [
      {
        "type": "ollama",
        "url": "http://127.0.0.1:11434",
        "ollamaName": "deepseek-r1:32b"
      }
    ],
    "reasoning": {
      "type": "tokens",
      "beginToken": "<think>",
      "endToken": "</think>"
    }
  }
]`

[iChristGit]

How it looks on production (hf.co/chat)

[iChristGit]

Also tried llama cpp and fails to websearch, seems like there is formatting issue

@nsarrazin can you help?