[Feature Request]: Option to disable reasoning/chain-of-thought in RAGFlow Chat Settings

[meetcharan-practice]

Self Checks

• [x] I have searched for existing issues search for existing issues, including closed ones.
• [x] I confirm that I am using English to submit this report (Language Policy).
• [x] Non-english title submitions will be closed directly ( 非英文标题的提交将会被直接关闭 ) (Language Policy).
• [x] Please do not modify this template :) and fill in all the required fields.

Is your feature request related to a problem?

When using Azure OpenAI LLM models inside RAGFlow’s Chat Settings, the model sometimes returns internal reasoning text such as:

“We need to answer what is ABC using the knowledge base…”
“First, I will analyze the context…”

These appear before the actual answer, even though the model is expected to output only the final response.

This behavior is not seen when using other LLM providers (e.g., LM Studio models), and it negatively affects the chat experience for end users.

Describe the feature you'd like

Please add a configuration option such as:

“Hide Chain-of-Thought / Reasoning” “Enforce Direct Answer Output Only” “Disable Reasoning Prefill for Azure OpenAI Models”

This option should ensure that: The model outputs only the final assistant answer Any reasoning, chain-of-thought, planning, or system-prefill steps are automatically suppressed RAGFlow enforces standard Azure OpenAI guidelines (no reasoning leakage)

Describe implementation you've considered

Why this is important:

Ensures clean, user-facing responses Prevents confusing or technical reasoning text from leaking Maintains consistency with how other LLMs behave

Documentation, adoption, use case

Additional Context:
This issue occurs regardless of the prompt structure. It seems related to how RAGFlow forwards system messages / context to Azure OpenAI. A toggle to strip reasoning would solve this.

Additional information

Model Name: STG-gpt-oss-120b

Related to https://github.com/infiniflow/ragflow/issues/11526