💸 noex-bot — AI Chatbot That Talks Like Les Gold "I'm Les Gold. I don't just talk — I close deals. And this bot? It talks like me!"

noex-bot is an AI-powered, voice-interactive chatbot that mimics the personality and voice of Les Gold from Hardcore Pawn. Using voice cloning, real-time speech recognition, and AI-generated responses, it gives you an animated, entertaining conversation experience — complete with a moving mouth.

🚀 Features 🧠 AI Chatting – Generates smart, witty responses using OpenAI's API.

🗣️ Voice Cloning – Uses F5 TTS with a fine-tuned Orkhon-TTS model to mimic Les Gold’s voice.

🧏 Speech Recognition – Captures your voice via Google Cloud Speech-to-Text.

👄 Mouth Animation – Pygame UI opens and closes the mouth while speaking.

🎤 Fully Voice-Driven – You speak, the bot listens and responds — no typing needed.

F5 TTS with a fine-tuned Orkhon TTS model (refer to the F5 TTS GitHub page for model setup) : https://github.com/SWivid/F5-TTS

Google Cloud credentials JSON for speech recognition

2. API Keys Create a .env file with the following:

OPENAI_API_KEY=your_openai_api_key
GOOGLE_APPLICATION_CREDENTIALS=credentials.json

Installation

First, create the environment

conda create -n noex python=3.10
conda activate noex

Then install the Pytorch for your cuda version

pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu128

Lastly, install the requirements using pip

pip install -r requirements.txt

Note! If you are using VSCode to run this application, don't forget to select the virtual environment you created as the interpreter or the program won't run properly!

🎬 Running the Bot

python start_conversation.py

The bot listens, thinks (via OpenAI), and replies in Les Gold’s voice.

Watch its mouth move as it talks.

🧠 Technologies Used OpenAI GPT (chat/completions)

Google Cloud Speech-to-Text

F5-TTS (Orkhon TTS fine-tuned)

Pygame for visual interaction

@article{chen-etal-2024-f5tts,
  title={F5-TTS: A Fairytaler that Fakes Fluent and Faithful Speech with Flow Matching},
  author={Yushen Chen and Zhikang Niu and Ziyang Ma and Keqi Deng and Chunhui Wang and Jian Zhao and Kai Yu and Xie Chen},
  journal={arXiv preprint arXiv:2410.06885},
  year={2024},
}

@misc{orkhon-tts,
  author = {Orkhon TTS Contributors},
  title = {Orkhon-TTS: Mongolic and Turkic multilingual TTS models},
  howpublished = {\url{https://github.com/talshaparov/orkhon-tts}},
  year = {2023},
}