Core Echo

About the Project

Core Echo is an advanced voice-driven AI agent orchestration system. It acts as a central hub that listens to user commands via voice, processes the audio, and dynamically routes or creates agents to handle the requested tasks. It features a Human-In-The-Loop (HITL) safety mechanism for destructive actions and a sandbox environment for executing plugins securely.

Use Case

This project is designed for developers or power users who want a hands-free, voice-activated AI assistant capable of performing a wide range of tasks locally. It can dynamically generate new agents on the fly if it doesn't have a pre-existing one for a specific request. The built-in web dashboard provides real-time visibility into the system's state, listening status, and agent execution.

Project Setup

1. Clone the repository and navigate into the project directory.
2. It is highly recommended to create and activate a Python virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows use `venv\Scripts\activate`

3. Install the required dependencies:

   pip install -r requirements.txt

   (Note: This project relies on local models and ML libraries such as torch, whisper-cpp-python, llama-cpp-python, and qdrant-client.)

Running the Application

To start the Core Echo assistant and the dashboard server:

python main.py

This will start listening to your voice input and simultaneously launch a local web dashboard (typically accessible at http://127.0.0.1:8000).