Implementation:Infiniflow Ragflow DialogService Save

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

Overview

Concrete tool for creating chat application (Dialog) records provided by RAGFlow DialogService.

Description

DialogService.save creates a new Dialog record via the inherited CommonService.save method. The POST /v1/dialog/set endpoint validates inputs (prompt_config presence, KB existence, LLM model availability), sets defaults for LLM settings and retrieval parameters, then persists the Dialog.

Usage

Called via the REST API to create or update a chat application configuration.

Code Reference

Source Location

• Repository: ragflow
• File: api/db/services/dialog_service.py (save: L54-67), api/apps/dialog_app.py (endpoint: L34-144)

Signature

class DialogService(CommonService):
    model = Dialog

    @classmethod
    def save(cls, **kwargs) -> Dialog:
        """Create a new Dialog record.

        Key kwargs:
            name: str - Dialog name.
            tenant_id: str - Owner tenant ID.
            kb_ids: list[str] - Linked knowledge base IDs.
            llm_id: str - LLM model ID.
            llm_setting: dict - {temperature, top_p, frequency_penalty,
                                  presence_penalty, max_tokens}.
            prompt_config: dict - {system, prologue, parameters, empty_response}.
            similarity_threshold: float - Default 0.2.
            vector_similarity_weight: float - Default 0.3.
            top_n: int - Default 6.
            top_k: int - Default 1024.
        Returns:
            Dialog model instance.
        """

Import

from api.db.services.dialog_service import DialogService

I/O Contract

Inputs

Name	Type	Required	Description
name	str	Yes	Chat application name
tenant_id	str	Yes	Owner tenant ID
kb_ids	list[str]	Yes	Knowledge base IDs to link
llm_id	str	No	LLM model (defaults to tenant default)
llm_setting	dict	No	Generation parameters
prompt_config	dict	Yes	System prompt and configuration
similarity_threshold	float	No	Default 0.2
vector_similarity_weight	float	No	Default 0.3

Outputs

Name	Type	Description
dialog	Dialog	Created Dialog model instance with generated ID

Usage Examples

import requests

url = "http://localhost:9380/v1/dialog/set"
payload = {
    "name": "Customer Support Bot",
    "kb_ids": ["kb-uuid-1"],
    "llm_id": "gpt-4-turbo",
    "llm_setting": {"temperature": 0.1, "max_tokens": 512},
    "prompt_config": {
        "system": "You are a helpful customer support assistant.",
        "prologue": "Hello! How can I help you today?",
        "empty_response": "I couldn't find relevant information."
    }
}
headers = {"Authorization": "Bearer <token>"}
response = requests.post(url, json=payload, headers=headers)

Related Pages

Implements Principle

• Principle:Infiniflow_Ragflow_Chat_Application_Creation