QUERYWISE

AI-powered Multi-Agent RAG system for intelligent document querying with fact verification

---

Overview

Querywise is a multi-agent Retrieval-Augmented Generation (RAG) system designed to help users query long, complex documents with accurate, fact-verified answers.
Unlike traditional chatbots like ChatGPT or DeepSeek, which may hallucinate responses or struggle with structured data, Querywise retrieves, verifies, and corrects answers before presenting them.

Key Features

• ✅ Multi-Agent System – A Research Agent generates answers, while a Verification Agent fact-checks responses.
• ✅ Hybrid Retrieval – Combines BM25 and vector search for more relevant content discovery.
• ✅ Multi-Document Handling – Dynamically selects the most relevant document, even with multiple uploads.
• ✅ Scope Detection – Prevents hallucinations by rejecting irrelevant queries.
• ✅ Fact Verification – Ensures responses are grounded in the source material.
• ✅ Web Interface with Gradio – Simple document upload and Q&A experience.

---

How Querywise Works

System Architecture

1️⃣ Query Processing & Scope Analysis

• Users upload documents and ask a question.
• Querywise checks query relevance and determines if the question is within scope.
• If the query is irrelevant, Querywise rejects it instead of hallucinating an answer.

2️⃣ Multi-Agent Research & Retrieval

• Docling parses documents into a structured format (Markdown, JSON).
• LangChain & ChromaDB power hybrid retrieval (BM25 + vector embeddings).
• With multiple documents uploaded, Querywise still identifies the most relevant sections.

3️⃣ Answer Generation & Verification

• The Research Agent generates a candidate answer.
• The Verification Agent cross-checks the response against source documents.
• If verification fails, a self-correction loop is triggered for refinement.

4️⃣ Response Finalization

• If verified, the final answer is presented to the user.
• If out of scope, Querywise informs the user instead of fabricating.

---

Why QUERYWISE Instead of ChatGPT or DeepSeek?

Feature	ChatGPT/DeepSeek ❌	Querywise ✅
Retrieves from uploaded documents	❌ No	✅ Yes
Handles multiple documents	❌ No	✅ Yes
Extracts structured data from PDFs	❌ No	✅ Yes
Prevents hallucinations	❌ No	✅ Yes
Fact-checks answers	❌ No	✅ Yes
Detects out-of-scope queries	❌ No	✅ Yes

Querywise is built for enterprise-grade document intelligence, research, and compliance workflows.

---

Installation

Quick Start

# Clone the repository
git clone https://github.com/Kaleemullah-Younas/Querywise.git
cd Querywise

# Install dependencies
pip install -r requirements.txt

# Set up environment variables
cp .env  # Edit with your API keys

# Run the application
python app.py

Environment Configuration

Create a .env file with the following variables:

OPENAI_API_KEY=your_openai_api_key
# Other optional configurations

---

Usage

1. Start the application with python app.py
2. Open your browser to http://localhost:7860
3. Upload one or more documents (PDF, DOCX, TXT supported)
4. Ask questions about your documents

---