https://github.com/inventwithdean/aurenia

Aurenia is a blazing fast native desktop study assistant
https://github.com/inventwithdean/aurenia

desktop-app lancedb llamacpp paddleocr rag rust tauri

Last synced: about 2 months ago
JSON representation

Aurenia is a blazing fast native desktop study assistant

• Host: GitHub
• URL: https://github.com/inventwithdean/aurenia
• Owner: inventwithdean
• License: apache-2.0
• Created: 2025-08-04T17:39:36.000Z (2 months ago)
• Default Branch: main
• Last Pushed: 2025-08-05T15:51:39.000Z (2 months ago)
• Last Synced: 2025-08-05T16:31:01.167Z (2 months ago)
• Topics: desktop-app, lancedb, llamacpp, paddleocr, rag, rust, tauri
• Language: JavaScript
• Homepage:
• Size: 32.1 MB
• Stars: 0
• Watchers: 0
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          


  

  
Aurenia

  
Offline, multilingual, beautifully yours.





  

  

  



---

> Aurenia is a blazing fast native desktop study assistant supporting top 25 global languages natively with features like OCR and RAG, that runs completely offline.






  



## The Vision

In a world where AI is becoming synonymous with the cloud, Aurenia challenges the trade-off between intelligence and privacy. It brings a powerful AI tutor to your desktop that runs entirely on your machine, ensuring your data is always secure.

## Key Features

* **🧠 Intelligent RAG Chat:** Ask complex questions about your documents. Aurenia's custom, multi-step RAG pipeline finds the precise information you need.

* **🌐 Truly Multilingual:** Have conversations, get translations, and generate summaries in your native tongue, with support for top 25 global languages.

* **📝 Interactive Study Tools:** Go beyond passive reading. Instantly generate interactive multiple-choice quizzes from any page to test your understanding.

* **🔒 100% Local & Private:** All AI processing happens on your device. Your documents and chats never leave your computer. No internet connection required.

* **👀 OCR for All Documents:** A built-in OCR engine automatically makes scanned documents and images fully searchable and interactive.

* **✅ No External Dependencies:** Everything comes packaged in the app. Just install, add the models, and you're ready to go.

* **💻 Minimal System Requirements:** Runs efficiently on as little as 8GB of RAM without a dedicated graphics card, and utilizes CUDA when available.

## Tech Stack

Aurenia is built with a modern, performance-focused stack:

| Component | Technology |

| :--- | :--- |

| **Application Framework** | [`Tauri`](https://v2.tauri.app/) (Rust + JS) |

| **AI Inference Engine** | [`llama.cpp`](https://github.com/ggml-org/llama.cpp) |

| **LLM Model** | Google's [`Gemma 3n`](https://huggingface.co/google/gemma-3n-E4B-it/tree/main) |

| **Embedding Model** | [`multilingual-e5-large`](https://huggingface.co/intfloat/multilingual-e5-large) |

| **Vector Database** | [`LanceDB`](https://github.com/lancedb/lancedb) |

| **OCR** | PaddleOCR (via [`paddle-ocr-rs`](https://github.com/mg-chao/paddle-ocr-rs)) |

| **PDF Handling** | [`pdf.js`](https://github.com/mozilla/pdf.js) |

## Installation & Usage

Getting started with Aurenia is easy:

1.  **Download the Installer:** Go to the [**Latest Release**](https://github.com/inventwithdean/aurenia/releases/latest) page and download the `.msi` file for Windows.

2.  **Download the GGUF Models (Prefer Quantized):** You need two model files to run Aurenia.

    * Download the LLM: [`Gemma 3n`](https://huggingface.co/unsloth/gemma-3n-E4B-it-GGUF)

    * Download the Embedding Model: [`multilingual-e5-large`](https://huggingface.co/phate334/multilingual-e5-large-gguf)

3.  **Place Models in Directory:** After installing Aurenia, place the two `.gguf` files you downloaded into the application's installation directory.

4.  Rename the `Gemma-3n`'s gguf to `model.gguf` and `multilingual-e5-large`'s gguf to `emb_model.gguf`

5.  **Launch Aurenia:** That's it! You can now open any PDF and start studying.

## Find Out More

* **▶️ Watch the Full Video Demo:** [https://youtu.be/n_-dwJi9wO8]

* **📄 Read the Technical Blog Post:** [https://inventwithdean.github.io/blog/aurenia/]

## License

This project is licensed under the Apache 2.0 - see the `LICENSE` file for details.

## Acknowledgements

This project would not be possible without the incredible open-source communities behind `llama.cpp`, `Tauri`, `LanceDB`, `pdfjs`, `PaddleOCR`, `intfloat/multilingual-e5-large` and the researchers at Google who developed and open-sourced the Gemma-3n models.