Lyrie Agent positions itself as the first autonomous AI agent with integrated cybersecurity, designed to act as a secure digital guardian. It targets users seeking robust protection alongside AI operational capabilities, offering a foundational security approach rather than an afterthought. The primary benefit is enabling users to run AI-powered operations with enhanced digital safety.

How It Works

Lyrie employs a four-layer architecture: an Interface layer (CLI, Web, Desktop, Mobile), an Agent Engine for core AI operations (spawning, skills, routing), a Memory Core featuring self-healing, versioned, and searchable Vector, Graph, and Structured memory, and crucially, "The Shield" layer. This Shield provides native, real-time cybersecurity including threat detection, anti-malware, WAF, behavioral analysis, and device protection, forming the project's core differentiator. It supports model-agnostic routing to various LLMs for optimized task performance.

Quick Start & Requirements

Installation can be initiated via a bash script: curl -fsSL https://lyrie.ai/install.sh | bash. Alternatively, manual setup involves cloning the repository (git clone https://github.com/lyrie-ai/lyrie-agent.git), navigating into the directory (cd lyrie-agent), installing dependencies with pnpm install, and starting the agent with pnpm start. Key prerequisites include Node.js and the pnpm package manager.

Highlighted Details

• Built with native cybersecurity as a foundational element, differentiating from platforms treating security as secondary.
• Features self-healing memory and self-improving skills for enhanced agent resilience and capability evolution.
• Supports intelligent, model-agnostic routing to diverse LLMs including Claude Opus, Grok, GPT-5.4, Gemini, Haiku, Qwen, and Gemma.
• Offers multi-channel connectivity via Telegram, WhatsApp, Discord, Slack, Signal, and CLI.
• Part of the "Lyrie World Pack" ecosystem, integrating with Lyrie Shield (device security) and Lyrie Mobile (companion apps).
• Provides a one-command migration tool (lyrie migrate --from openclaw) for users transitioning from OpenClaw.

Maintenance & Community

The project is developed by OTT Cybersecurity LLC. Specific details regarding community channels (e.g., Discord, Slack), active contributors, or a public roadmap are not detailed in the provided README.

Licensing & Compatibility

Lyrie Agent is released under the MIT License, permitting broad usage, forking, and building upon the software. This license generally supports commercial use and integration into closed-source projects without significant copyleft restrictions.

Limitations & Caveats

The README does not explicitly detail limitations, alpha/beta status, known bugs, or unsupported platforms. The codebase size is noted as <30K lines, significantly smaller than comparable projects like OpenClaw (430K lines), which may imply a narrower feature set or less mature implementation in areas outside core security and agent functionality.