ICI Core is a modular framework for building context, style, personality, and security-aware AI assistants. It targets developers and power users seeking to create personalized AI agents that can ingest data from various sources and respond intelligently. The framework offers flexibility in model selection, data storage, and customization, aiming to provide a robust foundation for advanced AI assistant development.

How It Works

The system operates through two primary pipelines: an Ingestion Pipeline for processing and storing data from sources like Telegram, Twitter, and YouTube, and a Query Pipeline for handling user interactions. It leverages vector databases for efficient context retrieval and allows for customizable response styles and personalities via prompt templates. Security is addressed through configurable input validation rules.

Quick Start & Requirements

• Installation: The recommended installation involves cloning the repository, running a setup.sh (macOS/Linux) or setup.bat (Windows) script, activating a virtual environment, and configuring a .env file with API keys. An experimental one-line installation is also available.
• Prerequisites: Python 3.10+, pip, Git, and NodeJS 18+. WhatsApp integration requires running a separate Node.js service.
• Configuration: Requires API keys for Telegram (API ID, API Hash) and a generator API key (OpenAI, Anthropic).
• Documentation: Comprehensive documentation is available in the docs directory.

Highlighted Details

• Supports multiple data sources including Telegram, Twitter, and YouTube.
• Offers flexible model selection with support for OpenAI, Anthropic, and Ollama providers.
• Features modular components with well-defined interfaces for extensibility.
• Includes configurable vector storage backends and security validation rules.

Maintenance & Community

The project is open-source with a contribution guide provided. Links to community channels or roadmaps are not explicitly mentioned in the README.

Licensing & Compatibility

Licensed under the MIT License, permitting commercial use and integration with closed-source projects.

Limitations & Caveats

The one-line installation is noted as experimental. Windows support for some installation options is also experimental. Specific details on supported vector databases or performance benchmarks are not provided.