Overview
Ailice is an open-source, general-purpose autonomous agent framework that uses the Interactive Agents Call Tree (IACT) to decompose complex tasks into composable agents and skills. It supports local and remote model integrations, multimodal I/O, and extensible modules.
Key features
- IACT architecture for fault-tolerant task decomposition and multi-agent orchestration.
- Support for multiple LLMs (local and cloud), inference services, and hybrid setups.
- Multimodal UI with voice, images, code highlighting, and file upload/download.
- Modular extension system and example notebooks for rapid development.
Use cases
- Research and education for multi-agent experiments and interactive notebooks.
- On-premises or privacy-sensitive deployments using local LLMs.
- Automation and system management via agent-executed workflows.
Technical details
- Implemented in Python, installable via pip and runnable with Docker; configuration-driven modules (config.json) for extension and MCP service integration.
- MIT licensed with active community, docs, and an online demo at the project site.