A real-time speech-to-speech chatbot, Weebo, addresses the need for seamless conversational AI interactions. It targets users seeking an AI assistant that can understand spoken input and respond verbally, leveraging advanced open-source models for speech recognition, language understanding, and text-to-speech synthesis. The primary benefit is a fluid, interactive conversational experience powered by local models.

How It Works

Weebo processes spoken input using Whisper Small, optimized via MLX for efficient, continuous speech recognition. Natural language understanding and response generation are handled by Llama 3.2. The generated text is then synthesized into speech in real-time using Kokoro-82M, supporting streaming response generation for a more natural conversational flow and offering support for different voices.

Quick Start & Requirements

• Installation: Requires downloading the kokoro-v0_19.onnx TTS model and pulling the llama3.2 model via Ollama. On macOS, espeak-ng must be installed via Homebrew, and ESPEAK_DATA_PATH exported.
• Primary Run Command: bash uv run --python 3.12 --with-requirements requirements.txt main.py. An alternative setup involves creating a virtual environment (python3 -m venv .venv, source .venv/bin/activate), installing requirements (pip install -r requirements.txt), and running python main.py.
• Prerequisites: Python 3.12, Ollama, Homebrew (for macOS), espeak-ng (for macOS).
• Documentation: A "Learn more here" link is mentioned in the README, but the specific URL is not provided.

Highlighted Details

• Real-time, continuous speech recognition via Whisper MLX.
• Natural language interaction powered by Llama 3.2.
• Real-time text-to-speech synthesis using Kokoro-82M.
• Supports streaming response generation for a more natural conversational feel.

Maintenance & Community

No specific details regarding contributors, sponsorships, community channels (Discord/Slack), or a public roadmap are provided in the README.

Licensing & Compatibility

The README does not specify a license type or any compatibility notes for commercial use or integration with closed-source projects.

Limitations & Caveats

The project explicitly states it "Works on Apple Silicon," indicating potential platform limitations. No other explicit limitations, known bugs, or alpha/beta status are mentioned.