Welcome to Theraxus AI - Your cutting-edge conversational AI companion, blending simplicity with powerful features to deliver an experience like no other! Whether it's chatting via text or voice, managing your documents, or using advanced Retrieval-Augmented Generation (RAG) capabilities, Theraxus AI has you covered. Dive in and explore a seamless, intelligent assistant designed to be highly functional, customizable, and super cool! 😎🌟
Because Theraxus isn't just an AI - it's YOUR AI. Built to run locally and supercharge your life, it's designed to keep you connected, informed, and entertained, all while offering powerful customization for developers and enthusiasts. Whether you want a text buddy, an audio companion, or a knowledge retriever, Theraxus is built to adapt to your needs. 💡
-
🗣️ Speech-to-Text (STT): Effortlessly convert spoken words to text with OpenAI's Whisper. Talk naturally; Theraxus listens. 🎤
-
🗨️ Text-to-Speech (TTS): Want to hear your assistant's voice? With pyttsx3, Theraxus AI can read things back to you smoothly. 🗣️
-
Multi-User Profile: Seamlessly switch and manage multiple user identities 👥 for a personalized and dynamic AI experience! 🔄
-
📚 Document Management:
- Upload Documents: Keep your files handy - add and manage text-based documents effortlessly.
- View Summaries: Get quick, concise summaries of your uploaded files.
-
🔍 Retrieval-Augmented Generation (RAG): Turbocharge your assistant's responses by intelligently retrieving the most relevant content from your document repository. Talk to your files like never before. ⚙️📖
-
📝 Chat History: Keep track of your past conversations with dedicated chat histories, so you never miss a beat. 📜
-
🔄 Modular and Customizable: Everything's designed in modules for easy customization. Add features, tweak behaviors, or integrate new models - make Theraxus truly yours! 🔧🛠️
-
Multi-Lingual Voice & Chat Support: Since the project is only limited by the model you use, you can do essentially a lot of things and take this next level as well.
Start by grabbing Theraxus AI:
git clone https://github.com/yourusername/theraxus cd theraxus
(Recommended) Create a virtual environment to keep everything neat:
python -m venv venv
Activate it with:
- Windows:
venv\Scripts\activate - macOS/Linux:
source venv/bin/activate
Theraxus relies on some system-level libraries:
- PortAudio: Install using Chocolatey (
choco install portaudio) for Windows, or Homebrew (brew install portaudio) for macOS. Use APT (sudo apt-get install portaudio19-dev) for Linux. - libsndfile: Same install path as PortAudio for each system.
Now, install Python dependencies:
pip install -r requirements.txt
Local AI model and Whisper models are needed for full functionality, I recommend GGUF models from hugging face, pick your choice:
- Place LLaMA model (e.g.
.gguf) in themodels/directory. - Whisper models download automatically when you start using STT.
You can rut it via gui.py or run individual modes by running the files below:
Choose between:
- Text-Based Chat:
python runllm.py - Voice-Based Chat:
python voice_runllm.py
Directory Structure
Theraxus/
├── config.py
├── database_manager.py
├── gui.py
├── rag_optimizer.py
├── runllm.py
├── stt.py
├── tts.py
├── voice_runllm.py
├── requirements.txt
├── README.md
├── models/
│ └── Your_Model.gguf
├── data/
│ ├── audio/
│ ├── docs/
│ ├── conversations/
│ ├── vector_db/
│ │ ├── index/
│ │ └── embeddings/
│ └── cache/
│ ├── whisper_models/
│ └── transformers_models/
├── logs/
│ └── theraxus.log
└── .gitignore
Welcome to Theraxus AI! Type 'exit' to quit.
You: load
Processing files... Added 'document1.txt'. Added 'document2.pdf
You: Explain this file to me
AI: Based on your input, here's a summary...
Voice Interface Ready! Speak into your microphone. Listening... You: load [File selection dialog opens] You: Sumaarize the file for me AI: Based on your input, here's a summary...
Theraxus AI is built for YOU to customize! 🛠️ Add more TTS/STT features, swap Models without breaking the code choose, Mistral, Gemma, LLama or Any model for richer responses, expand the RAG capabilities, build Vision Capabilities or extend API capabilities, build bigger projects - the sky's the limit! 🌌. Since the project is modular, a lot of settings can be changed from config file only, no more time wasted in changing deeper elements into the main code.
Change configurations in config.py:
STT_CONFIG = { 'MODEL_NAME': 'small', # Modify to 'small' or 'medium' for Whisper. 'SAMPLE_RATE': 16000 }
TTS_CONFIG = { 'RATE': 180, # Faster speech rate 'VOLUME': 0.8, # Lower volume }
Logs are stored in logs/theraxus.log 📂, where you can monitor everything happening under the hood. 🛠️
Sample Log:
2024-04-27 15:30:45,123:INFO:Created new database file: data\conversations\conversations.json 2024-04-27 15:30:45,789:INFO:Generated response for input: Hello
I'd love for you to contribute! Make Theraxus AI even better. Fork it, add features, or report issues - let's build something amazing together! 🛠️
How to Contribute:
- Fork the Project 🍴
- Create a Feature Branch (
git checkout -b feature/AmazingFeature) 🌱 - Commit your Changes (
git commit -m 'Add some AmazingFeature') 💬 - Push to the Branch (
git push origin feature/AmazingFeature) 📤 - Open a Pull Request 📨
Distributed under the Apache License. See LICENSE for more information.
📝 Note: Please remember to credit Theraxus AI whenever you use or build on its capabilities. Sharing knowledge is caring! 🤗
Big shout-out to:
- OpenAI Whisper for making STT possible. 🎙️
- pyttsx3 for TTS. 🗣️
- Llama for Llama.cpp
- Hugging Face for models
- hnswlib for efficient document retrieval. 📚
- Sentence-Transformers for meaningful sentence embeddings. 💡
Hey there! Just a quick heads-up: This minimal version of Theraxus AI was developed by me alone, so I've removed a lot of the more intricate code to avoid any clashes with other systems. What you'll find here is the core structure and all the main features, but simplified to ensure it's easy to understand and integrate. I assume you know what you're doing and have all understanding of modification and implementation.
Think of this as a foundation---a light version that still captures the essence of what Theraxus is all about, but without the bells and whistles that might complicate things. It's great to get you started, and you can always build upon it to make it your own!
Please remember to provide credit when using this version, in line with the Apache License 2.0. Thanks for understanding and enjoy exploring Theraxus.
Let's make your conversations smarter, your documents more accessible, and your AI experience truly awesome! 🥳