🦊🔗 FurChain

🌟 Create Lifelike Digital Personas 🌟

Empower your digital experiences with lifelike characters, complete with their own voices and personalities. FurChain is your toolkit for creating and interacting with digital personas that feel real.

🚀 Features

• 🎤 Voice Cloning: Instantly clone any voice and generate speech with unparalleled realism.
• 🎭 Effortless & Rapid Role-Play Creation: Instantly bring to life endless role-play scenarios with minimal effort and maximum speed.
• 🛠️ Offline and Open Source: Full control and customization without the need for an internet connection.

🌐 Examples

• 🎶 Music Voice Alteration: Change the voice in a song on-the-fly while preserving the original background music. Check examples/singer_change.py.
• 🤖 Self-Clone Chatbots: Craft your own chatbot with a cloned voice and character in mere seconds. See examples/clone_me.py for a demonstration.
• 🎲 Text-Based RPGs: Kickstart a text RPG adventure swiftly. Experience it in examples/adventure_with_fox.py.

📋 Requirements

To ensure the best experience with FurChain, your system should meet the following requirements:

• CPU: X86_64 architecture
• GPU: Nvidia GPU (recommend >= 12GB VRAM)
• Docker Compose: For managing multi-container Docker applications.
• NVIDIA Container Toolkit: For GPU support within Docker containers.

🏃‍♂️ How to Run

Starting the API Service

Run the following command to launch the service:

docker compose up -d

This spins up 5 containers, each serving a unique purpose:

• text-mongo: Manages chat history with MongoDB, optimized for CPU usage.
• text-llama-cpp: Hosts the Llama-cpp server for large language models, utilizing GPU.
• audio-gpt-sovits: Runs the GPT-SoVITS server for voice synthesis, leveraging GPU.
• audio-funasr: Powers the FunASR online server for speech recognition, optimized for CPU usage.
• audio-rvc: Operates the RVC server for voice conversion, utilizing GPU.

Initiating FurChain

You can use docker to have a quick glance at the capabilities of FurChain:

docker run --rm --gpus all --network host -v $PWD/data:/app/data -it markyfsun/furchain python3 examples/adventure_with_fox.py

The data directory serves as the central hub for your models, audio files, and chat logs:

• gpt-sovits: Holds reference audio for speech generation.
• llama-cpp: Contains LLM GGUF files available for download from huggingface.
• rvc: Stores models for voice conversion.

Should you wish to contribute or customize furchain, clone this repository and install it as an editable package:

git clone https://github.com/furchain/FurChain.git
cd FurChain
pip install -e .