guard-rag 1.0.9

Privacy-first, fully offline AI document assistant secured by tiered safety guardrails

     

GuardRAG

A privacy-first, fully offline AI document assistant — secured by a tiered safety guardrails system

Upload any document. Ask anything. Get answers — entirely on your machine.
No cloud. No API keys. No data leaves your device.

📦 Installation · 📖 Documentation · 🤝 Contribute

---

🚀 Overview

GuardRAG is a powerful, production-ready command-line tool that lets you chat with your documents offline. By combining LangChain, Ollama, and FAISS, it provides a private alternative to cloud-based RAG solutions. It includes a built-in safety engine that protects sensitive data using a tiered sensitivity system.

---

🛠 Why GuardRAG?

Most RAG chatbots rely on cloud APIs, which creates privacy risks for sensitive documents — contracts, medical records, internal reports. GuardRAG solves that by:

• Local Inference: Runs models locally via Ollama.
• Offline Embeddings: Uses HuggingFace transformers strictly on your device.
• Tiered Safety: 4 levels of guardrails (Public → Restricted).
• Pro Design: Clean, modern CLI and Web interfaces.

---

✨ Features

• 100% Offline — No network calls at runtime.
• Multi-format Support — PDF, TXT, DOCX.
• Persistent Memory — Disk-cached FAISS indexes for rapid re-queries.
• Privacy-First Guardrails — Integrated protection against jailbreaks and PII leaks.
• Built-in Web UI — Optional browser-based interface included in the package.

---

⚙️ Data Sensitivity Tiers

Level	Badge	Protection Scope
Public		Jailbreak & prompt injection detection.
Internal		+ API keys, credentials, tokens.
Confidential		+ SSN, email, phone, DOB, credit cards.
Restricted		+ Medical, HIPAA/GDPR, financial data.

---

📥 Installation

Install GuardRAG directly from PyPI:

pip install guard-rag

Prerequisites:

• Ollama installed and running.
• Pull a model (e.g., ollama pull gemma3:1b).

---

🏁 Quick Start

Once installed, simply run the guard-rag command:

1. Simple Interactive Chat

guard-rag --pdf my_doc.pdf

2. Custom Model & Sensitivity

guard-rag --pdf secret_report.pdf --model llama3.1 --sensitivity Confidential

3. Launch Web Interface

# Running without arguments launches the built-in Web UI
guard-rag

This will start an interactive chat session with your document.

CLI Options

guard-rag --pdf <file>             Load and chat with a PDF document
          --model <model>          Ollama model to use (default: gemma3:1b)
          --ollama-host <url>      Ollama server URL (default: http://localhost:11434)
          --chunk-size <int>       Document chunk size (default: 1000)
          --chunk-overlap <int>    Chunk overlap (default: 200)
          --sensitivity <level>    Data sensitivity: Public | Internal | Confidential | Restricted
          --no-guardrails          Disable safety guardrails
          --help                   Show this help message

Example Session

# Start with a PDF using Llama 3.1
guard-rag --pdf report.pdf --model llama3.1 --sensitivity Confidential

# You: What are the key findings?
# Chatbot: Based on the document, the key findings are...

---

Project Structure

GUADRAILS-RAG-CHAT-TOOL/
│
├── guardrag/                 # Main installable package
│   ├── api/                  # FastAPI local server
│   ├── cli/                  # Command-line interface
│   ├── rag/                  # RAG pipeline logic
│   └── utils/                # General utilities
│
├── docs/                     # Documentation (INSTALL, QUICK_REFERENCE)
├── tests/                    # Unit and integration tests
├── scripts/                  # Development and maintenance scripts
├── extras/                   # Experimental / legacy components
│
├── pyproject.toml             # Modern build configuration
├── setup.py                   # Legacy support configuration
├── README.md                  # Project overview
├── CONTRIBUTING.md            # Contribution guidelines
├── CODE_OF_CONDUCT.md         # Community standards
└── LICENSE                    # MIT License open source

.guardrag_storage/ is auto-generated on first document load (FAISS cache).

---

Configuration

Environment Variables

Copy .env.example to .env and adjust as needed:

cp .env.example .env

Variable	Default	Description
OLLAMA_HOST	http://localhost:11434	Ollama API endpoint
NO_PROXY	huggingface.co,...	Bypass proxy for local+HF calls
PORT	8000	Server port (auto-set by PaaS)

Chunking Parameters

Adjustable per-session via the sidebar in the UI:

• Chunk Size (default 1000 chars)
• Chunk Overlap (default 200 chars)

Different chunk settings for the same file produce a separate FAISS index automatically.

---

Deployment

From PyPI (recommended)

pip install guard-rag

From Source

git clone https://github.com/sowmiyan-s/GUADRAILS-RAG-CHAT-TOOL.git
cd GUADRAILS-RAG-CHAT-TOOL
pip install .

In a virtual environment (best practice)

python -m venv .venv
# Windows:
.venv\Scripts\activate
# macOS / Linux:
source .venv/bin/activate

pip install guard-rag

---

Contributing

Contributions are welcome! Please see CONTRIBUTING.md for details on our code of conduct and the process for submitting pull requests.

Bug reports and feature requests are welcome via GitHub Issues.

---

License

This project is licensed under the MIT License — see LICENSE for details.

---

Built with ❤️ by Sowmiyan S

FastAPI · LangChain · Ollama · HuggingFace · FAISS · Vanilla JS