Implementation:Infiniflow Ragflow Async Ask

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

Overview

Concrete tool for executing search queries with SSE-streamed AI answers and related features provided by RAGFlow's conversation endpoints.

Description

The POST /conversation/ask endpoint streams AI-synthesized search answers via SSE using async_ask. Companion endpoints POST /conversation/mindmap and POST /conversation/related_questions provide structured knowledge visualization and follow-up suggestions.

Usage

Called by the search application frontend via the useSendQuestion hook.

Code Reference

Source Location

• Repository: ragflow
• File: api/apps/conversation_app.py (L394-423 ask, L426-441 mindmap, L444-478 related_questions)

Signature

@manager.route("/ask", methods=["POST"])
@login_required
@validate_request("question", "kb_ids")
async def ask_about():
    """Search with SSE-streamed AI answer.
    Body: {question: str, kb_ids: list[str], search_id?: str}
    SSE: data:{code:0, data:{answer, reference}}
    """

@manager.route("/mindmap", methods=["POST"])
@login_required
@validate_request("question", "kb_ids")
async def mindmap():
    """Generate mind map from search results.
    Body: {question: str, kb_ids: list[str], search_id?: str}
    Returns: Mind map dict.
    """

@manager.route("/related_questions", methods=["POST"])
@login_required
@validate_request("question")
async def related_questions():
    """Generate related search questions.
    Body: {question: str, search_id?: str}
    Returns: list[str] of suggested questions.
    """

Import

from api.db.services.dialog_service import async_ask, gen_mindmap

I/O Contract

Inputs

Name	Type	Required	Description
question	str	Yes	Search query
kb_ids	list[str]	Yes	Knowledge base IDs
search_id	str	No	Search app ID (loads search_config)

Outputs

Name	Type	Description
SSE stream	text/event-stream	data:{code:0, data:{answer, reference}} chunks
mindmap	dict	Structured mind map
related_questions	list[str]	Suggested follow-up queries

Usage Examples

import requests

# Search with SSE streaming
url = "http://localhost:9380/conversation/ask"
payload = {"question": "What is RAGFlow?", "kb_ids": ["kb-uuid-1"]}
with requests.post(url, json=payload, headers=headers, stream=True) as r:
    for line in r.iter_lines():
        if line.startswith(b"data:"):
            print(line.decode())

Related Pages

Implements Principle

• Principle:Infiniflow_Ragflow_Search_Query_Execution