Superagentsys/novalclaw

What it solves

OmniNova Claw provides a unified, local-first platform for creating and managing AI agents. It solves the fragmentation of AI tools by combining multiple LLM providers, various communication channels (like Slack and Discord), and a complex memory system into a single desktop control plane and runtime.

How it works

Built with a high-performance Rust core and a Tauri/React desktop interface, the system uses a modular architecture:

• Soul System: Uses MBTI psychology to define agent personalities, reasoning patterns, and communication styles.
• Three-Layer Memory: Employs Working Memory (short-term context), Episodic Memory (long-term interaction history), and Semantic Memory (knowledge bases from skills and documents).
• Skills & Tools: Implements an extensible system where capabilities are loaded via SKILL.md files, including built-in tools for web search, file operations, and shell execution.
• Connectivity: A gateway manages routing between various LLM providers (OpenAI, Anthropic, etc.) and deployment channels.
• Mobile Integration: Dedicated iOS and Android clients enable phone-based agent capabilities, such as automated call screening and spam detection.

Who it’s for

It is designed for developers and power users who want to build sophisticated, personalized AI agents with local control, cross-platform accessibility, and the ability to deploy bots across multiple messaging platforms.

Highlights

• MBTI-Driven Personas: Architect agent cognition based on psychological types for distinct behavioral patterns.
• Multi-Channel Deployment: Native support for Slack, Discord, Telegram, WeChat, WhatsApp, and more.
• Local-First Runtime: High-performance Rust core that can run headlessly on servers or as a desktop app.
• Mobile Phone Agent: Integrated iOS and Android apps for call handling, ASR/TTS, and on-device spam detection.
• Extensible Skill System: Ability to import specialized domain knowledge and tools via standardized skill packs.