NGIBS - Next Generation Intelligent Browsing System

NGIBS is a privacy-first, AI-powered desktop research assistant built with Python, PyQt6, and Ollama. It helps users search the web, gather relevant information, and generate structured answers while keeping the workflow local and controllable.

Unlike many cloud-based AI tools, NGIBS is designed to work with locally hosted models and local memory, making it a strong choice for users who care about privacy, offline access, and full control over their data.

Screenshots

Here are a few examples of the interface in action:

Why NGIBS?

NGIBS combines the power of modern large language models with web retrieval, local memory, and a desktop interface. It is especially useful for:

• Researchers who need fast, citation-aware answers
• Developers exploring technical topics and documentation
• Students gathering information from multiple sources
• Privacy-conscious users who prefer local AI tools

Key Features

1. Quick Search

Use the language model for fast answers without depending on web search.

• Fast responses
• Works with local models
• Good for general knowledge and short tasks
• Uses session-based memory for continuity

2. Live Search

Retrieve up-to-date information from the web and combine it with AI reasoning.

• Web search using DuckDuckGo
• Wikipedia support
• Web content extraction with BeautifulSoup
• Retrieval-based answers with citations
• Helpful for recent or dynamic topics

3. Deep Search

Go beyond simple retrieval with recursive reasoning and multi-step analysis.

• Multi-step research workflow
• Better context handling
• Useful for complex questions
• Supports exporting results as PDF, Markdown, or DOCX

4. Context-Aware Memory

Keep track of conversations and important context over time.

• Short-term session memory
• Long-term vector-based memory
• Multi-chat support
• Persistent conversation tracking

Additional Capabilities

• File attachment support
• Model management from inside the app
• Local model download and removal
• Memory management tools
• Privacy-focused local-first architecture
• Desktop app packaging support with PyInstaller

Tech Stack

NGIBS is built using the following technologies:

• Python
• PyQt6 for the desktop UI
• Ollama for local LLM execution
• PyTorch and Transformers
• ChromaDB for vector memory
• BeautifulSoup for content parsing
• DuckDuckGo and Wikipedia APIs for retrieval
• Markdown, PDF, and DOCX export support
• PyInstaller for packaging

Privacy First

NGIBS follows a local-first design philosophy:

• No forced cloud dependency
• Models run locally when available
• Users control which models are installed
• Data remains under user control
• Conversation memory is stored locally

Supported Platform

• ✅ Windows (current focus)
• 🔜 macOS
• 🔜 Linux

Installation (Windows)

One-Click Installer

For the simplest setup, download install_ngibs.bat from the repository and double-click it. The launcher will use the checkout beside the batch file when available; otherwise it will create %LOCALAPPDATA%\NGIBS\app and download the repository there.

It checks each component before doing work and only installs what is missing:

• Python 3.10 or newer (Python 3.11.9 is downloaded when no suitable version exists)
• Ollama and its local service
• The default llama3.1:latest model
• A project-local .venv and the packages in requirements.txt

The launcher can be run repeatedly. It does not recreate the virtual environment, reinstall already-satisfied pip packages, or pull an existing model. It uses Git when available and falls back to the public GitHub ZIP archive when Git is not installed.

The batch file's configuration values are at the top of install_ngibs.bat: REPO_URL, INSTALL_ROOT, PYTHON_DOWNLOAD_VERSION, PYTHON_MIN_MAJOR, PYTHON_MIN_MINOR, OLLAMA_MODEL, and OLLAMA_PORT. Change OLLAMA_MODEL there if the application should use a different default model; keep it consistent with backend/engine.py.

Prerequisites

No manual Python, Ollama, or package installation is required. The computer needs Windows, an internet connection for the first setup, and permission to install per-user applications. Git is optional because the launcher has a ZIP fallback.

Setup

git clone https://github.com/avarshvir/NGIBS.git
cd NGIBS
python -m venv .venv
.venv\Scripts\activate
pip install -r requirements.txt

Download a Model

After installing Ollama, pull a model such as:

ollama pull llama3.1:latest

You can also use other supported models depending on your hardware and preferences.

Run the Application

python main.py

How to Use NGIBS

• Start the app and choose your preferred AI model
• Use Quick Search for fast general answers
• Use Live Search for current web information
• Use Deep Search for more detailed research workflows
• Manage memory and chat sessions from the built-in controls

Project Structure

• main.py: desktop application entry point
• backend/: core engine, runtime, search tools, memory, and storage logic
• screenshots/: example UI screenshots

Upcoming Features

Planned future improvements include:

• PyWebView integration
• Image and video search support
• Multimodal AI features
• Voice input and voice output
• Plugin-based extensions
• Browser-like tab previews
• Knowledge base building tools
• Research graph visualization
• Export to Notion and Obsidian

Contribution

Contributions, bug reports, ideas, and feature requests are welcome. If you would like to improve the project, feel free to open an issue or submit a pull request.

---

Made by Arshvir, AI/ML Engineer | Open-Source Contributor | Indie Developer :)