Implementation:Infiniflow Ragflow SearchService Update LLM Config

Knowledge Sources	RAGFlow
Domains	RAG, Search
Last Updated	2026-02-12 06:00 GMT

Overview

Concrete tool for configuring LLM model and generation parameters for search answer synthesis provided by RAGFlow SearchService.

Description

Uses the same POST /search/update endpoint with search_config.chat_id (LLM model) and search_config.llm_setting (generation parameters). The frontend validates with a Zod schema.

Usage

Called via the search settings form in the frontend.

Code Reference

Source Location

• Repository: ragflow
• File: api/apps/search_app.py (L64-118), web/src/pages/next-search/search-setting.tsx (L65-103)

Signature

# Same endpoint as Search Configuration
# search_config.chat_id = LLM model ID
# search_config.llm_setting = {temperature, top_p, frequency_penalty, presence_penalty}

Import

from api.db.services.search_service import SearchService

I/O Contract

Inputs

Name	Type	Required	Description
search_config.chat_id	str	No	LLM model ID
search_config.llm_setting	dict	No	Generation parameters

Outputs

Name	Type	Description
search_app	dict	Updated search application

Usage Examples

import requests

url = "http://localhost:9380/search/update"
payload = {
    "search_id": "search-uuid-123",
    "name": "Product Search",
    "search_config": {
        "chat_id": "gpt-4-turbo",
        "llm_setting": {
            "temperature": 0.3,
            "top_p": 0.9
        }
    }
}
requests.post(url, json=payload, headers=headers)

Related Pages

Implements Principle

• Principle:Infiniflow_Ragflow_Search_LLM_Settings