-
Notifications
You must be signed in to change notification settings - Fork 1
Home
This project is an AI virtual assistant chatbot that uses speech recognition, text-to-speech, and voice activation detection to have a conversation with you. It is locally hosted and uses an open source LLM using LM Studio, faster-whisper for speech-to-text, and openai-edge-tts for text-to-speech and Flask for the web interface in a browser.
The project was a way for me to get back into learning coding again by learning how to use Python and studying the general functions and usage of LLMs within NLP. The project has slowly become a hybrid study on the use and understanding of the English language just as much as it is the study of LLMs.
I will update this welcome home page with latest updates and patch notes, as well as current and future planned features, as well as other things as the project grows.
- Speech & Text Chat
- Locally Hosted LLM
- Memory Management System
- User Preference Learning
- Interest Tracking
- Mood Detection & Context
- Thinking Module with Inner Thoughts
- Modular Architecture
- And More!
Major update featuring complete modular architecture, enhanced chat functionality, and comprehensive AI intelligence systems.
- Modular File Organization - Complete restructuring for better maintainability
- Enhanced Chat Module - Updated API integration and improved chat history handling
- Memory Management System - Three-tier storage for preferences, interests, and memories
- Advanced Preferences Module - Hybrid regex/NLP preference extraction with confidence scoring
- Interest Tracking System - Dynamic interest detection with sentiment analysis and decay
- Mood Detection & Context - Hybrid regex/transformer mood detection with dynamic context
- Thinking Module - Inner thoughts generation and cognitive processing foundation
📝Click here to read the full patch notes.
BnuuyBotApp v0.5.7 features a comprehensive modular architecture with dedicated packages for all major AI systems. The core functionality is stable and includes:
✅ Speech-to-text and text-to-speech capabilities
✅ Web-based chat interface with Flask
✅ Locally hosted LLM integration via LM Studio
✅ Memory management with three-tier storage system
✅ Preference extraction with confidence scoring
✅ Interest tracking with sentiment analysis and decay
✅ Mood detection with hybrid regex/transformer approach
✅ Thinking module for inner thoughts and cognitive processing
✅ Comprehensive modular architecture with clean encapsulation
✅ Enhanced chat system with improved API integration
Follow these steps to get BunnyBot up and running.
-
Install Prerequisites:
- Install Python 3.10+, LM Studio, and Docker.
- Set up a local server in LM Studio and load an LLM of your choice.
- Run the Docker quick start setup for TTS functionality.
-
Clone the Repository:
git clone https://github.com/LumiBunny/BnuuyBotApp.git
-
Install Python Dependencies:
pip install -r requirements.txt
- Run the Application:
python app.py
As this is an app in early development, I have not used or tested it on any other hardware. Here is my current PC configuration:
📝 See my full PC configuration, including drivers and software, on the Wiki for more details.
These features are planned for future development:
- Voice Activated Commands: Voice commands to control the app, similar to voice assistants
- Enhanced Sentiment Analysis: Real-time sentiment grading for both user and chatbot responses, with potential visual feedback for animated character integration
- Advanced Memory Functions: Voice-controlled memory operations (remember, search, update, delete), reminders, timers, and note-taking
- Conversation Intelligence: Smarter context understanding using the integrated memory, preference, and interest systems