QwenKB MVP

Powered by Alibaba Cloud DashScope · Connecting...

Create Knowledge Base

📚

Loading knowledge bases...

Upload Documents

Supported: PDF, DOCX, TXT, MD, CSV. Documents are automatically chunked, vectorized, and stored.

📂

Click or drag files here to upload

Max 50MB per file

Query Knowledge Base

Agent API Reference

External agents can query knowledge bases via these endpoints. All endpoints return JSON.

List Knowledge Bases

GET/api/agent/kb

Query (RAG or Search)

POST/api/agent/query
Form parameters:
kb_id (required) — Knowledge base ID
query (required) — Your question
top_k (optional, default 5) — Number of chunks to retrieve
mode (optional, default "rag") — "rag" for full answer, "search" for raw results

Example: curl

curl -X POST http://localhost:8000/api/agent/query \
  -F "kb_id=YOUR_KB_ID" \
  -F "query=What is the company policy?" \
  -F "mode=rag"

Example: Python

import requests resp = requests.post("http://localhost:8000/api/agent/query", data={ "kb_id": "YOUR_KB_ID", "query": "What is the company policy?", "mode": "rag" }) print(resp.json()) # {"answer": "...", "sources": [...], "chunks_used": 5}

All Endpoints

GET/api/kb — List knowledge bases
POST/api/kb — Create knowledge base (JSON: name, description)
DELETE/api/kb/{id} — Delete knowledge base
POST/api/kb/{id}/documents — Upload document (multipart)
GET/api/kb/{id}/documents — List documents
DELETE/api/documents/{id} — Delete document
POST/api/kb/{id}/query — RAG query (JSON: query, top_k)
POST/api/kb/{id}/search — Vector search (JSON: query, top_k)