An Advanced AI Research Assistant Agent

📖 Overview

DeepQuery is a professional-grade AI Research Assistant designed to process, analyze, and synthesize complex information from multiple sources. Utilizing Retrieval-Augmented Generation (RAG), it seamlessly handles multi-document uploads, web searches, and image processing while maintaining both short-term context and long-term memory.

Engineered with token efficiency and safety in mind, DeepQuery doesn't just summarize—it critically analyzes data, detects conflicting information across sources, and provides precise citations for every claim.

✨ Key Features

🧠 Advanced AI & RAG Capabilities

• Intelligent Memory: Utilizes both short-term conversational context and long-term memory retrieval.
• Multi-Modal Processing: Supports concurrent analysis of multiple document formats (PDFs, Excel, etc.) and images.
• Web Search Integration: Enhances localized knowledge with real-time web context.
• Conflict Detection: Automatically cross-references multiple documents and flags conflicting information between sources.
• Precise Citations: Every piece of synthesized knowledge is tracked back to its specific document or web source.
• Token Efficiency: Highly optimized prompt engineering and a "safety buffer" mechanism prevent token waste and context window choking.

🛠️ Specialized Research Tools

DeepQuery offers granular control over the AI's behavior via specialized chat modes:

• Temporary Chats: Disable long-term memory and storage for sensitive or one-off inquiries.
• Strict Quote Mode: Forces the agent to answer using direct, verbatim quotes from documents/websites without altering the original context.
• Challenge Mode: The AI critically analyzes its own synthesis to identify gaps, biases, or missing perspectives.
• Audience Targeting: Toggle explanations between General or Expert levels.
• Reasoning Depth: Adjust processing depth (Low, Medium, High) depending on whether you need a quick summary or a deep analytical breakdown.

⚙️ Powerful Admin

Built with Django Jazzmin, the admin panel provides complete operational control:

• LLM & Embedding Management: Easily toggle and configure local models or API providers (OpenAI, Anthropic, Gemini, HuggingFace).
• User Activity: Monitor chats, messages, and system usage.
• Subscription & Billing: Manage subscription tiers, daily limits (messages, images, documents), and process payments via the Zarinpal SDK.

🏗️ Enterprise-Grade Architecture

• Clean & Decoupled: Apps are heavily decoupled based on domain logic, ensuring near-zero coupling.
• Auto-Registry Factory Pattern: Easily plug-and-play new LLMs or embedding providers without touching core logic.
• Asynchronous Processing: Celery and Redis handle heavy embedding tasks and background jobs (like chat title generation) to keep the UI perfectly responsive.
• Robust Security: Secured via JWT, CORS, and strict Django permission classes. Includes unit tests for mission-critical components.
• Test Client: Includes a lightweight Vanilla JS frontend for easy API testing and demonstration.

---

📸 Screenshots

Main Interface	Image Upload Usage

Source Conflict Detection	Admin Dashboard

---

💻 Tech Stack

Core Framework & API:

• Python 3.x
• Django 6.0.3 & Django REST Framework 3.17.1
• SimpleJWT, drf-spectacular (Swagger/OpenAPI)
• Django-Jazzmin (Admin UI)

Database & Task Queue:

• PostgreSQL + pgvector (Vector Search)
• Redis 7.4 & Celery 5.6

AI & LLM Integration:

• OpenAI, Anthropic, Google GenAI, HuggingFace Hub
• Tiktoken

Data Processing:

• Pandas, Numpy, PDFPlumber, PyPDFium2
• BeautifulSoup4, Openpyxl

---

🚀 Getting Started

Prerequisites

• Python 3.10+
• PostgreSQL (with pgvector extension installed)
• Redis Server

Installation

1. Clone the repository:

git clone https://github.com/aminrezaeeyan/DeepQuery.git
cd DeepQuery/server

2. Set up a virtual environment:

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies:

pip install -r requirements.txt

4. Environment Variables: Create a .env file in the root directory and add your configurations (Database credentials, API Keys, Redis URL, Zarinpal tokens, etc.).

5. Database Setup:

python manage.py makemigrations
python manage.py migrate
python manage.py createsuperuser

6. Run Background Workers (in a separate terminal):

celery -A deepquery worker -l info

7. Start the Development Server:

python manage.py runserver

🛡️ Testing

Run the included test suite to verify core components:

pytest

🤝 Contributing

Contributions, issues, and feature requests are welcome!

📄 License

This project is licensed under the MIT License.