GitHub
FreeFully autonomous, general-purpose AI agent
FreeFree tier
Inputs: textOutputs: text, code
About GitHub
AIlice is a fully autonomous, general-purpose AI agent that aims to replicate the functionality of a JARVIS-like assistant using open-source large language models. It features the Interactive Agents Call Tree (IACT) architecture, enabling decomposition of complex tasks into dynamically constructed sub-agents with high fault tolerance. AIlice can perform thematic research, coding, system management, literature reviews, and complex hybrid tasks. It supports local LLM execution, voice dialogue (via ChatTTS), and MCP tools. The project is open-source and emphasizes self-evolution of AI agents.
Key Features
Interactive Agents Call Tree (IACT) architecture for dynamic task decomposition
Supports local open-source LLMs as well as commercial models (Claude 3.5/3.7, Gemini 2.5 Pro)
Voice dialogue functionality powered by ChatTTS
Integration with MCP (Model Context Protocol) tools
Capable of thematic research, coding, system management, and literature reviews
Self-evolution capability for autonomous feature expansion
Online demo available at kragent.ai
High fault tolerance through recursive agent construction
Pros & Cons
Pros
- Fully autonomous operation with no manual intervention required
- Open-source and free to use, with no licensing restrictions
- Can run locally with open-source models, ensuring privacy
- Uses innovative IACT architecture for robust task decomposition
- Actively maintained with frequent updates and community support
- Supports voice interaction for a more natural user experience
Cons
- Requires technical expertise to set up and configure locally
- Performance heavily depends on the underlying LLM quality
- Some features (e.g., voice dialogue) are still experimental
- May require significant computational resources for local models
- Limited documentation for advanced customization
Best For
Conducting thematic research and literature reviewsAutomating coding tasks and software developmentSystem administration and managementHandling complex hybrid tasks that combine multiple domainsBuilding and expanding AI agent capabilities autonomously