Alise - The AI Agent

Alise is a terminal-based AI agent designed to be modular and extensible. It aims to assist with various tasks, including coding, system administration, and more, by leveraging different AI providers and local capabilities.

Current Features (Initial Implementation)

• Core Agent Logic (src/core/agent.js):
  • Supports basic supervised and unsupervised modes (currently placeholders).
  • Initializes and manages the agent's lifecycle.
• Memory System (src/core/memory.js):
  • Persistent memory using a memory.json file in the project root.
  • Stores session data, API keys, and other relevant information.
• Modular Architecture (src/core/moduleLoader.js, src/modules/):
  • A basic module loading system that can load JavaScript modules from the src/modules directory.
  • Modules can extend Alise's functionality (demonstrated with sampleModule.js).
• API Key Management (src/utils/apiKeyManager.js):
  • Utilities to add, view (obscured), retrieve, and delete API keys for different AI providers.
  • Keys are stored securely using the agent's memory system.

Project Structure

• index.js: Main entry point for the application.
• package.json: Project metadata and dependencies.
• memory.json: Stores persistent data for the agent (e.g., API keys).
• src/: Contains the core source code.
  • core/: Core components like the agent logic, memory, and module loader.
  • modules/: Directory for pluggable modules that extend Alise's capabilities.
  • utils/: Utility scripts, such as the API key manager.

Getting Started (Conceptual)

Currently, Alise is in its very early stages of development. To run the agent (for development purposes):

1. Ensure you have Node.js installed.
2. Clone the repository.
3. Install dependencies (if any beyond Node.js built-ins are added, e.g., npm install).
4. Run the agent: node index.js

The index.js file currently contains example code to demonstrate the implemented features.

Future Development Ideas

• Full implementation of supervised and unsupervised modes.
• Integration with AI providers (Anthropic, Google A2A).
• Advanced memory capabilities (e.g., vector stores, RAG).
• Interactive CLI for user interaction.
• Self-improvement capabilities (editing its own code).
• Robust error handling and logging.